New Challenges in Cross-Domain Deterrence

Home , Collective punishment, Deterrence theory, Limited war, Tactical victory

Perspective

C O R P O R A T I O N Expert insights on a timely policy issue

King Mallory

fter a period of U.S. primacy that followed the end of that have long relied on the Pax Americana—extended American the Cold War, the United States has been confronted deterrence of aggression against them—to guarantee both national with successful actions on the part of Russia and China and regional security.2 to revise the territorial status quo in Ukraine and the At the same time that the use of hybrid and terrorist tactics AWest Pacific. Both countries employed “gray zone” or “hybrid war- of warfare has gained newfound salience in the land domain of fare” tactics in pursuing these goals. After its 2001 and 2003 inva- warfare, the probability that future military conflict will encompass sions of Afghanistan and Iraq, the United States was challenged by conflict in space and cyberspace has risen significantly. Not only a significant increase in activity on the part of transnational groups has the United States’ ability to deter aggression in the traditional of nonstate actors employing terrorist tactics of warfare as well. The air, land, and sea domains of warfare been cast in doubt, but new cumulative activities of all of these actors have cast in doubt the requirements to deter future aggression in the domains of space and territorial status quo in Europe, the Middle East, North and sub- cyberspace have also arisen. When an opponent has no incentive Saharan Africa, and South and East Asia. to initiate or escalate conflict at any given intervention or escala- All sets of actors have employed asymmetric military tactics. tion threshold in any given domain of warfare—both vertically These tactics have been designed to avoid direct conventional and horizontally within that domain and laterally into one or more military confrontation with the United States in areas of warfare in additional domains of warfare—successful cross-domain deterrence which the United States dominates and has superior power projec- can be said to be in effect. tion capabilities.1 These developments have unsettled traditional This Perspective examines ways and means by which the U.S. allies in Europe, Asia, and the Middle East and North Africa United States and its allies might meet these new challenges in cross-domain deterrence. It first situates deterrence within the Figure 1. Alternative Strategy Sets in the broader spectrum of strategies available to international actors Strategies Spectrum when pursuing their vital interests. Definitions both of deterrence Preempt and of different types of deterrence are elaborated, and key assump- Control Prevent tions and enabling factors for successful deterrence identified by Compel classic texts are summarized. Changes in the world system since the Coerce Punish classic texts on deterrence were written are noted, and the need for Deter Deny and definition of cross-domain deterrence are elaborated. Induce Dissuade Cross-domain deterrence in four discrete domains or subareas Reassure

of warfare is then examined: space, hybrid warfare, terrorism, and Degree of force or coercion Conciliate Cooperate cyberwarfare. In each case, the functioning of the classical enablers Accommodate is scrutinized, and possible remedial measures are suggested. Poten- Strategies and measures tial strategies of deterrence by threat of denial of the benefits sought and of deterrence by threat of punishment are then suggested. SOURCES: Based on data from Schelling, 1966; G. H. Snyder, 1961; Freedman, Strategies that can be implemented within the given domain or 1981; Freedman, 2004; and Huth, 1991. subarea are examined, as well as strategies that require action across one or more additional domains. The discussion closes by suggest- Conciliation, in contrast, involves removing key obstacles to reach- ing how to prioritize between competing deterrent strategies and by ing an agreement, without agreeing to a major part of the other highlighting a number of policy implications. side’s demands. If the number of concessions made to reach agreement is not excessively one-sided (at which point they may consti- The Spectrum of Strategies tute appeasement*), limited reciprocal concessions are a politically Nations have a spectrum of strategies and measures that they can legitimate option by which to avoid or terminate a conflict or reach employ to shape their relations with other nations and nonstate an agreement.4 actors. To clarify the definition and role of deterrence in overall Dissuasion comprises all forms of persuasion, including reassur- U.S. national security policy, the strategy sets that are available are ance and inducement, that would cause an adversary not to follow briefly summarized in this section; they are cooperation, dissua- a particular course of action. Reassurance involves measures to allay sion, coercion, and control (see Figure 1).3 an opponent’s concerns by convincing the opponent that a situation Cooperation involves working together in pursuit of a common goal. Accommodation means agreeing to a substantial but relatively * Professional jargon or terms of art from the deterrence literature are enclosed in painless portion of the other side’s demands to achieve agreement. quotation marks or italicized when first introduced.

2 is less threatening or more benign than originally thought. Induce- rence by punishment aims to make a conflict too painful or danger- ment involves sweetening the pot by providing incentives to reach ous and thereby coerce the opponent into avoiding or terminating an outcome.5 it. All-out punishment can be incompatible with attempts to coerce Control involves the deliberate use of force (political, economic, an enemy to make a desired decision: It is difficult to influence an or military) to restrict an adversary’s strategic choices; it depends aggressor when it has nothing left to lose.11 Deterrence by denial originally on judgments about the opponent’s strategy, but those is coercive in part but essentially tends toward threats to control judgments eventually become irrelevant as the adversary runs out the situation sufficiently to deny the adversary strategic options or of options. Lawrence Freedman distinguishes between two types gains.12 As a general proposition, whenever feasible, deterrence by of controlling strategy: prevention and preemption.6 Prevention denial is to be preferred to deterrence by punishment because the involves an actor exploiting its existing strategic advantages to latter requires continuous coercion, whereas the former involves deprive an adversary of the capability to pose a threat before that control.13 In addition to deterrence by denial and deterrence by threat has become imminent; it deals with problems before they punishment, at least four additional different types of deterrence become crises.7 Preemption involves forestalling losses from an can be distinguished; they are neither mutually exclusive nor mutu- opponent first strike that is believed to be imminent.8 ally exhaustive: Coercion uses threats of force to influence an opponent’s 1. General deterrence is said to be in effect when the balance strategic choices. For coercion to succeed, the opponent must be of power is stable and no actor is considering mounting an able to choose the path of compromise. There are two types of attack on another. General deterrence can be in effect at the coercion. Compellence involves persuading an adversary that it must global level or at a regional level. act for fear of the consequences if it does not. A compellent threat 2. Immediate deterrence is required when an actor starts to is intended to persuade the opponent to give up something of value. contemplate or prepare for military action, thereby unleashing It is a strategy designed to make others act in ways they consider a crisis or emergency and causing general deterrence to break harmful to themselves but that benefit the compellor.9 This Per- down.14 spective focuses on deterrence, the other type of coercion. It involves 3. Direct deterrence, also known as central deterrence, involves a threats to force a potential opponent into forgoing a possible course deterrer threatening a potential aggressor with retaliation to of action. It is a policy that seeks to persuade an adversary, through prevent the aggressor from using military force against the the threat (implicit or explicit) of retaliation, that the expected costs deterrer’s most vital interests, such as its homeland. Because of initiating or continuing the use of coercion or military force to direct deterrence involves the defense of vital interests, it is resolve a conflict will outweigh the expected benefits. It operates generally believed to involve a credible threat.15 both before and during a conflict, either by punishing the adversary 4. Extended deterrence involves a deterrer threatening retalia- or by denying it the benefits sought through its aggression.10 Deter- tion against a potential aggressor in an attempt to prevent

3 the potential aggressor from pursuing a certain course of Table 1. Deterrence: Classical Assumptions and action against an ally (or protégé). Because extended deter- Enablers rence involves defending non-core interests of another state, Assumptions Underlying Enablers of Effective the probability that the deterrer will actually carry out the Deterrence Deterrence retaliation threatened is regarded as lower than in the case Shared normative framework Salience of the deterrent threat(s) of direct deterrence, in which a deterrer is defending its own Antithetical interests Clarity of the deterrent threat(s) 16 vital interests. Valuables that can be placed Timeliness of the threat(s) at risk Key Contributors to Successful Deterrence Risk sensitivity or, at least, risk Credibility of the threat(s) Effective deterrence is far from easy to achieve. Analysis of classical neutrality z Reputation of the party making the threat(s) texts on deterrence theory indicates that for a strategy of deterrence z Legitimacy or propor- to succeed, in addition to being clear, timely, and credible, a num- tionality of the threats(s) ber of further assumptions must be met, and enablers (summarized Limited rationality Contribution of technology to in Table 1) must be present. Many of the factors initially identified stability have been debated and emended in the subsequent literature. A Clarity of escalation thresholds shared normative framework and interests that are not diametri- Ability to counter threshold cally opposed (i.e., a zero-sum game) were basic assumptions that manipulation classical writers initially thought must hold.17 However, Patrick SOURCES: Schelling, 1966, pp. 236, 244; G. H. Snyder, 1961, pp. 10, 15, 19, 27, 48, 97–98, 99, 128, 168, 200, 209, 234; Huth, 1991, pp. 6, Morgan later recognized that a shared normative framework is not 9, 11, 30, 31, 33–34, 35 (note 13), 43, 50, 53, 54, 137–138, 200, 201, a requirement for deterrence.18 The party whom the deterrer seeks 203–204; Freedman, 2004, pp. 22, 33, 35–36, 49, 55; and Trager and to deter is also assumed to have something it values that the deter- Zagorcheva, 2005. rer can hold at risk. The two parties’ relative risk profiles matter: It assumed to act with limited rationality.20 For deterrent strategies is more difficult and costly to deter an opponent who has displayed to work, the potential aggressor must be aware of the deterrer’s risk-seeking behavior. It is not possible to deter an opponent who is threat and understand its logic. The greater the threat’s salience and totally insensitive to risk. It must be noted, however, that behavior clarity, the greater its potential credibility. A state’s reputation for that one party may subjectively consider to be risk-seeking may carrying out threats, as opposed to bluffing, matters. actually be the result of a sober, objective assessment on the part of Bluffing and then caving have a significant negative effect; they the other party.19 Strategies of deterrence also rely on the assump- lead potential aggressors to a markedly higher future estimation tion that the parties will decide and act rationally. But this “rational that the deterrer is bluffing when making deterrent threats. The actor” assumption too has been relaxed with time. Parties are now cost of reversing such a conclusion, once it has been formed in the

4 mind of the potential aggressor, is high. In the midst of a conflict, can escalate vertically to another, higher and more deadly level the cost of reversing such an impression can even be prohibitive.21 within a given domain of warfare (vertical escalation thresholds and/ The deterrer must be able to avoid both “strategic surprise” and or vertical escalation).26 “tactical surprise”22 and thereby have the time in which to carry Aggressors deliberately create ambiguity around crisis situ- out its deterrent threat(s) before the aggressor presents it with an ations to achieve their goals. A broad range of proxy actors is accomplished fact. The credibility of a strategy of deterrence cannot available to create the impression that acts of aggression are being be separated from the political objectives it is supposed to support; carried out by means beyond the control of and not attributable to they must be legitimate. For public audiences to consider a deter- the ultimate aggressor that actually instigated them.27 These strata- rent strategy to be politically legitimate, it is important that the gems are intended to sow confusion and uncertainty in the deter- measures that are threatened in response to opponent actions be rer’s ranks and to create a reasonable doubt as to the identity and perceived to be proportionate.23 responsibility of the ultimate instigator of the aggression. One goal The state of technology can either contribute to or detract from is to deny the international community and the deterrer adequate the effectiveness of a deterrent threat; it thereby affects both intra- warning and ability to identify the ultimate source of an act of conflict first-strike and crisis stability.24 Both a capability to retaliate aggression; another goal is to deny them adequate time to mobilize that can survive an opponent’s first strike (“first-strike stability”) the domestic and international political support needed to respond. and the requirement that a relatively high number of weapons be The overarching goals are thus to achieve strategic and tactical expended to eliminate any one element of the other side’s retalia- surprise and to delay and delegitimize as disproportionate any tory capacity (a high attacker-to-target ratio) militate in favor of organized response to the aggressor’s actions. To deter successfully, successful deterrence and strategic stability. During the Cold War, a the deterrer must be able to thwart such attempts to manipulate or significant reserve capacity orstrategic slack was built into the aggre- compromise its intervention and escalation thresholds.28 gate number of strategic weapon systems held. Successful deterrence was thereby ensured by creating an ability to “ride out” an opponent Cross-Domain Deterrence first strike, while retaining sufficient retaliatory capacity to inflict The world has changed since the principles of classical deterrence unacceptable damage on the other side.25 Although the concepts of theory, summarized above, were first elaborated in the late 1950s strategic slack and the attacker-to-target ratio were developed for and early 1960s. The United States’ position of overwhelming the purpose of nuclear warfighting, as explained below, they can be economic dominance has declined in relative terms. Europe has applied to other weapon systems and aspects of warfare as well. recovered to become a significant economic competitor. China has The thresholds that first trigger actions threatened as part of become the world’s largest economy. Beijing’s recent behavior sug- a deterrent strategy (intervention thresholds) are another important gests that it is bent on using its newfound power to restore a Sino- element. So are the intra-conflict break points at which violence centric security system in Asia, to challenge the post–World War II

5 territorial settlement in the West Pacific, and to revise the post-war graphic theater of military operations, military leaders and analysts international security architecture into one that reflects and accom- have increasingly chosen to highlight the need to deter potential modates a multi-polar world order.29 adversary aggression within and across all five domains of military This changing “correlation of forces”30 makes it increasingly activity (air, land, sea, space, and cyberspace).35 unlikely that the United States will be able to achieve its interna- Because of their greater recent salience, this Perspective focuses tional goals by acting alone. In its international dealings, Washing- on the new challenges of deterring aggression in three of those five ton will likely be compelled by circumstance to abandon thoughts domains: of primacy and to revert to a modernized form of the “grand 1. space strategy”31 of collective security that served the United States and 2. land—the focus here is on two subareas of land combat: its allies well for almost 50 years, from the end of World War II to • repeated employment of hybrid warfare tactics by potential the end of the Cold War. adversaries Within this broader context, the United States and its allies • continuing aggression by nonstate actors employing terror- must decide on the limited set of potential conflicts that they ist tactics can reasonably expect to be able to deter. Among a total of eight 3. cyberspace. strategic goals, the 2015 U.S. National Security Strategy identifies countering asymmetric terrorist tactics EFUFSSJOHBHHSFTTJPOCZ What form of cross-domain escalation might one hypotheti- 3VTTJB and deterring aggression CZ China.32 A doctrine of cross- cally need to deter in this new landscape? The central column in domain deterrence might reasonably be limited to these three goals Figure 2 represents the two domains that are outside the focus of alone, given the complexity and allied resource requirements this paper (air and sea). The two columns to the left represent the involved in such a doctrine’s implementation.33 subareas of hybrid warfare and nonstate actors in the land domain. During the Cold War, military strategists primarily focused The two columns to the right represent the newly salient space and on deterrence of a Warsaw Pact conventional or nuclear attack that cyberspace domains. Within each domain are notional escalation would take place in Europe on land, in the air, and at sea. These thresholds at which conflict can be intensified to a more violent differing domains of military operations were largely conflated level (indicated by the dotted red lines). Figure 2 shows a puta- and were understood to be included in the term theater of military tive path by which vertical escalation mostly takes place across, as operations. The potential for future conflict in space has become opposed to within, domains; it charts a nine-step cross-domain more salient since that time. In 2008, computers were first used escalation path.36 This notional escalation path does not take place as tools of aggression (cyberwarfare) in support of a conventional strictly following each of the thresholds in each of the domains but military war, the Russo-Georgian War.34 Because war in space and instead skips both across domains and over some of the escalation cyberspace cannot be limited to the boundaries of a single geo- thresholds within each domain:

6 Figure 2. A Cross-Domain Escalation Path 6. cyberattacks on enemy critical infrastructure 7. the destruction of U.S. early warning satellites 8. a preemptive special operations forces attack on theater weapons of mass destruction 9. nuclear weapons employment.37

In each domain or subdomain depicted, thresholds exist at which the United States or its allies might choose to fi rst intervene militarily or to escalate military activity vertically to a new, more Cyber warfare Space warfare Nonstate actors Hybrid warfare Vertical escalation Vertical intense level of violence. At each such threshold, the United States and its allies—and U.S. opponents—have the option of initiating Conventional or nuclear warfare or escalating military activity laterally into one or more additional domains of military activity. At each threshold, the participants Lateral escalation have the further option of escalating the confl ict horizontally by

Escalation thresholds Escalation path drawing one or more additional regions, countries, or nonstate actors into the confl ict. When an opponent has no incentive to NOTES: The central column represents conventional warfare and notional initiate or escalate confl ict at any given intervention or escalation escalation thresholds within the air and sea domains. The two columns to the right of the center column represent the new domains of space and cyberspace. The two threshold in any given domain of warfare—both vertically and columns to the left of the center column show notional escalation thresholds in the subareas of hybrid warfare and nonstate actors in the land domain of warfare. horizontally within that domain and laterally into one or more additional domains of warfare—successful cross-domain deterrence 1. a low-level cyber information operations (IO) or “trolling” can be said to be in eff ect.38 campaign Th e text that follows examines in detail each of the four focus 2. lateral and vertical movement to hybrid, cross-border actions areas described: space, hybrid warfare, terrorism, and cyberspace. by proxies In each case, the presence or absence of the contributors to suc- 3. lateral and vertical escalation to a state-sponsored terrorist cessful deterrence, identifi ed in classic texts and summarized attack in Table 1, is scrutinized, and possible remedial measures are 4. rising and crossing to “blind” U.S. satellites to prevent detec- suggested.39 Th e applicability of the various types of deterrence tion of mobilization described previously and the possibility of strategic and tactical 5. outbreak of conventional hostilities surprise are investigated. Technology’s infl uence in achieving successful deterrence is reviewed. Th e state of U.S. declaratory deter-

7 rence doctrine in the area in question is also assessed. Potential Types of deterrence: Because the balance of power in space is strategies of deterrence by threat of denial of the benefits sought being challenged by Russia and China with the implicit threat of a by the adversary and of deterrence by threat of punishment of the first strike, general deterrence in space can be said to be low, even opponent are then suggested. Strategies that can be implemented if the threat of opponent attack is not imminent. As it has demon- within the given domain or subarea are examined first. Deterrent strated its own ability to shoot down satellites in LEO, the United strategies that require action across one or more additional domains States has a medium-level capability for immediate deterrence in are examined next. The focus here is on deterrence of the initiation space. This capability is not high because the United States does of conflict and of the vertical and lateral escalation of a conflict that not appear to be able to shoot down GEO satellites or satellites in has already started. The more complex, but surmountable, chal- highly elliptical orbit (HEO). Because potential aggressors depend lenge of containing horizontal escalation risks is not dealt with in less on space for warfighting than the United States does, opportu- this Perspective. nities for direct deterrence appear to be low.42 As the United States is currently hard pressed to defend its own satellites, let alone those Space of others, opportunities for extended deterrence in space seem Enablers: Of all the domains of military operations examined limited as well.43 in this paper, the contributors toward successful deterrence identi- Surprise: Given the lift required to get to GEO (where such fied in the classic texts appear to be least present in space. China U.S. crown jewels as the Space-Based Infrared System and the demonstrated an ability to attack U.S. satellites in low Earth orbit Advanced Extremely High Frequency Nuclear Command and (LEO) and in geosynchronous Earth orbit (GEO) in 2007 and Control [NC2] satellites are located), it is unlikely that strategic 2013, respectively. Beijing demonstrated its ability to conduct surprise can be achieved by launching a sneak attack on these rendezvous and proximity operations with U.S. satellites, in 2016. assets. The infrared signature accompanying the launch of a missile Russia demonstrated similar capabilities in 2015 and 2016.40 Both fired for this purpose would probably be detected, and the mis- China and Russia have thus made it clear that they have the capa- sile’s trajectory could then be mapped. The same is not true of bility to carry out and may be contemplating crippling blows on air-launched antisatellite (ASAT) attacks on objects in LEO or of U.S. space-based assets at the outset of a conflict. Because of their attacks by maneuverable exo-atmospheric kill vehicles launched potentially devastating impact, the United States might be forced before the outbreak of a conflict. Because it can retaliate against to take strong countermeasures in reaction to such attacks. Given LEO satellites, the United States’ ability to avoid tactical surprise is this fact, a strike on U.S. space-based assets at the outset of a crisis not low, even if opponent GEO and HEO satellites may remain out may betray a high appetite for risk on the part of U.S. opponents. of reach. An opponent with a high appetite for risk is more difficult to Technology: In space, the attacker-to-target ratio refers to the deter.41 number of ASAT weapons required to kill an opponent satellite.

8 Strategic slack refers to the availability of a reserve satellite stockpile conclude that the short-term advantages and benefi ts expected from and of a capacity to surge launch such replacement satellites. In attacking U.S. space-based assets outweigh the expected costs. space, both the attacker-to-target ratio and strategic slack appear to One way of looking at the threat to U.S. and allied military be low: A single ASAT shot can take out a high-degree vertex in the satellites is to disaggregate those platforms’ functions and to exam- network of U.S. military satellites. Stocks of replacement satellites, ine which ones are most susceptible to attack and which forms of substitute capabilities, and surge launch capacity do not appear to attack are most eff ective. Figure 3 compares functions of satellites be high. (communication, reconnaissance, targeting, assistance in naviga- Doctrine: Although the United States has of late made it clear tion, surveillance, and NC2) against various methods or targets of that it will retaliate against attacks in space, the type and severity attack (dazzling of satellites with lasers, attempts to jam transmis- of attack that would elicit a response have not been specifi ed, nor sions, creating fi elds of debris in space that might damage satellites, has the kind of response that would ensue. Th ere is thus no fully permanently blinding satellites with lasers, destroying satellites articulated and widely disseminated strategy for deterring attacks with various types of kill vehicles, and disabling or destroying in space. Th e United States has not formally laid out strong “red one or more of the space-based components of the U.S. nuclear lines” for deterrence in space that might shape future norms for kill chain).45 Th e check marks indicate that the method of attack acceptable behavior by spacefaring nations. U.S. deterrent strategy in space therefore lacks both salience and clarity.44 Due to the fact Figure 3. Space Threat Matrix that U.S. statements concerning intervention thresholds remain fuzzy, the credibility and reputation of U.S. declaratory deterrence Seriousness of disruption or attack policy in space must be judged to be low. Nuclear Percentage It might be argued that, on its own, the objective fact that Dazzling Jamming Debris Blinding Destroying kill chain with Communication 50% core U.S. interests are at stake in space will deter opponents from a Reconnaissance 66% fi rst strike, regardless of U.S. doctrine. However, the United States Targeting 83% (1) depends on space-based assets for modern warfi ghting capabili- Navigation 66% Surveillance 83% Importance of function ties, (2) has failed to demonstrate its ability to continue to function NC2 50% with degraded support from space, and (3) has failed to identify Percentage with 50% 83% 66% 50% 100% 50% ensuing retaliatory punishment signifi cant enough to eliminate opponents’ considerable incentive to carry out a fi rst strike. Arguably, NOTES: The check marks indicate whether the type of disruption or attack indicated in each column can be applied to the function identified in each row. As in the absence of clarity and an indication of political will about the a rough guide, the percentage of functions that can be targeted by each type of kinds of retaliation that an aggressor may expect to encounter from attack is then calculated for each column (the x-axis percentages). Similarly, the percentage of attack modes to which each function is vulnerable is calculated for the United States, a sober-minded aggressor may therefore objectively each row (the y-axis percentages).

9 would apply to the function. The black lines arenotional escalation an attack on all. The aggressor would face the prospect of collective thresholds below or to the right of which the forms of attack or the retaliation. military satellite function put at risk by such an attack might be Cross-domain deterrence: As in the Cold War, U.S. and allied important enough to warrant a military response.46 The matrix gives armed forces can also deter attacks in space through patterns of a rough indication of the activities and actors of potentially greatest annual exercise and training behavior that demonstrate to potential concern in the bottom-right quadrant and suggests that satellite aggressors that they are increasingly able to function with degraded surveillance and targeting functions and the kinetic destruction support from space. Disaggregation of the functions carried out and jamming of satellites may be the greatest threats faced in space. by satellites of the kind shown in Figure 3 allows nonstrategic In-domain deterrence: In-domain deterrence of attacks in functions for which there are air-, land- or seaborne substitutes to space might be achieved by denying the opponent the benefits be identified with a view to off-loading some share of those func- sought. The wartime pooling of allied commercial and military tions from U.S. military satellites in the future. Exercises and the satellite services is a form of denial that could be used to expand off-loading of noncritical communications functions from satellites extended deterrence to space. Over a period of 15 to 20 years, the onto a connected Pacific Ocean seabed fiber optic network are both future topology of the United States’ network of military satellites examples of cross-domain deterrence by the threat of denial.49 might be shaped more proactively than it has been to date. The goal Cross-domain deterrence by threat of punishment consists of would be to create a connected network in which information flows retaliation designed to achieve a countervailing impact or effect in efficiently. A connected network will decay gracefully under attack, other domains equivalent to the one that the aggressor intended to thereby remediating the significant current risk that the network of achieve by attacking the deterrer in space. Kinetic or nonkinetic U.S. military satellites will fail catastrophically when subjected to attacks on adversary command, control, communication, intel- directed attack.47 Combining this reshaping of the network with a ligence, surveillance, and reconnaissance (C3ISR) and reconnais- more even distribution of capabilities across satellites and a surge sance, surveillance, targeting, and attack (RSTA) assets in the launch capacity (reducing the probability of tactical surprise) might land, air, and sea domains are ways of blinding the aggressor and make the space domain a contributor to crisis stability, rather than disorganizing its command and control. Such attacks would have a detractor from it.48 In space, in-domain deterrence by the threat an effect on the aggressor similar to that intended by an attack on of punishment might include a counterattack on the aggressor’s U.S. space-based assets. Kinetic attacks of this kind would cause military satellites. A capability to attack opponent satellites in HEO loss of life and would likely be considered escalatory by opponents. and GEO would boost the credibility of such a threat. An alterna- It is, however, in the U.S. national interest to increase the likelihood tive approach might be an international collective security agree- that adversaries conclude that retaliation of this kind is inevitable ment that considers an attack on one ally’s military satellite systems and therefore not intended to be escalatory. Doctrine and exercises could impress this point upon adversaries. In a turnaround play,

10 cross-domain punishment might also be achieved by threatening to attack adversary infrastructure in the land and cyber domains that The United States and its allies can deter is designed to ensure regime survival in the face of key long-term an opponent preemptive first strike on U.S. political vulnerabilities. space-based assets at the outset of a conflict Arguably, one of the greatest weaknesses of certain U.S. by threatening a response that would put the adversaries is that they lack true democratic political legitimacy and accountability. Because of this vulnerability, these opponents adversary regime’s long-term survival at risk seek to create protected national “information spaces”50 in which by destroying its control over its protected their government administration alone creates and controls the domestic information space. dominant political narrative disseminated by domestic mass media. The creation of such protected spaces prevents the widespread might enable objective facts at variance with opponent government dissemination of facts at variance with or contradictory of incum- narratives to be widely disseminated to adversary mass audiences. bent regime narratives. A protected information space prevents the In addition to potentially threatening long-term regime survival, dissemination of information about regime violations of the rule of providing such dissonant pieces of factual information to adversary law, corruption, nepotism, and incompetence that are potentially mass domestic television audiences can make it more difficult for threatening to long-term regime survival. The United States and its opponents to sustain, let alone dominate, the political narrative allies can exploit this weakness by mapping the network of instru- either domestically or internationally—that is, to win the informa- ments by which opponents create a protected information space51 tion war—during times of crisis.52 and threatening, in the event of conflict, to attack these assets either by cyberattack or with ordnance. The United States and its Hybrid Warfare allies can deter an opponent preemptive first strike on U.S. space- According to the North Atlantic Treaty Organization (NATO), based assets at the outset of a conflict by threatening a response “hybrid threats are those posed by adversaries, with the ability that would put the adversary regime’s long-term survival at risk by to simultaneously employ conventional and non-conventional destroying its control over its protected domestic information space. means adaptively in pursuit of their objectives.”53 The termnon- The United States might further deter attacks in space by conventional means is interpreted broadly here. Paramilitary and proactively penetrating the defenses of the adversaries’ protected covert activities include the infiltration of subversive operatives into information space. Modernizing a successful Cold War strategy, the zone of conflict, sabotage and the fomenting of rebellion, the resources can be focused and pooled to provide objective, factual provision of military materiel, the involvement of “volunteers” or round-the-clock television news programming directly from satel- “advisers” who provide training in the use of military equipment, lites into television set-top boxes in opponent countries. This action the involvement of such “volunteers” or “advisers” in actual combat,

11 and the direction of combat operations. The range of actors sup- lenge the balance of power or mount attacks on others. Immediate porting such activities extends from “troll armies” to hackers in the deterrence of hybrid warfare means that successful measures have cyber domain to military trainers and advisers, “volunteers,” covert been taken to prevent the further employment of such tactics after operators, proxy terrorist organizations,54 and special operations their first use. Hybrid warfare tactics were employed in the Russo- forces whose uniforms and equipment have been sanitized of means Georgian war of 2008. These tactics were then employed again dur- to identify their national origin (e.g., “little green men”). Acts of ing Russia’s March 2014 annexation of Crimea and its subsequent economic coercion involve making aid conditional on adopting interventions in East Ukraine. Similar tactics have been repeatedly political positions desired by the donor nation, making the provi- employed by China in affronting its neighbors in the East China sion of aid conditional on the recipient not accepting assistance Sea and the South China Sea. Both general and immediate deter- from the donor’s international rivals, repeated shutoffs of energy rence of the employment of hybrid warfare tactics can therefore be supplies in critical winter months or threats to do so, threats to said to be low. organize a selloff of the sovereign debt of adversary nations, and Direct deterrence of hybrid warfare is possible. However, direct threats to restructure reserve currency holdings in coordina- deterrence requires the deterrer to recognize that hybrid warfare tion with allied opponent nations. Acts of economic punishment tactics are being used; ignore the intimidation involved in the fact include product bans, an elevated frequency of customs inspec- that the deterrer is generally being preyed upon by a larger, more tions, border closures, bans of the export of key commodities, and powerful aggressor; and show the organization, determination, harassment of locally resident citizens of the target country by and political will to mount a response that will have an immediate immigration authorities.55 These are legal acts of harassment that deterrent effect on the aggressor. Most of the countries recently tar- Herman Kahn termed retortions.56 The range of actors supporting geted with hybrid warfare tactics (e.g., Georgia, Ukraine, Vietnam, or carrying out acts of economic coercion or punishment is broad and the Philippines) have, in one way or another, failed to measure and encompasses commercial fishing vessels or fishing fleets; state- up to these requirements. Other countries targeted (e.g., Japan owned enterprises (e.g., oil exploration platforms); sovereign wealth and Indonesia) are just beginning to satisfy them by rapidly and funds; state-owned banks; state-owned development banks; and robustly reacting to insults sustained. As a result, countries’ general maritime surveillance, fishery protection, and coast guard vessels.57 ability directly to deter hybrid warfare tactics can still said to be Enablers: The assumptions underlying deterrence theory low. largely seem to hold for the hybrid warfare subarea of the land Extended deterrence of hybrid warfare tactics involves prevent- domain. However, less than half of the factors identified as contrib- ing the use of hybrid warfare tactics against U.S. allies through uting to successful deterrence appear to be present.58 the threat of retaliatory measures. In both the South China Sea Types of deterrence: General deterrence of hybrid warfare can cases of Chinese poaching in the lagoon of Scarborough Shoal and be said to be in force when such tactics are not employed to chal- of Chinese territorial claims to the Senkaku Islands, the United

12 States successfully deterred Chinese use of hybrid warfare tactics.59 tation in deterring this form of warfare must currently be judged to However, these two successes have so far proved to be exceptions be low. rather than the rule. U.S. intervention failed to deter further In-domain deterrence: In-domain strategies of deterrence of continuing use of such tactics on China’s part. Dozens of other economic and paramilitary hybrid warfare by the threat of denial incidents involving China’s use of hybrid warfare tactics to assert of the benefits include heightened case-by-case scrutiny of trans- revisionist Western Pacific territorial claims remain unanswered. actions by adversary state-owned economic vehicles in Western Russia’s use of such tactics in Georgia, Crimea, and East Ukraine markets; denying state-owned economic vehicles access to Western remains unchallenged as well. Because U.S. intervention failed to markets unless equal, reciprocal, unrestricted access to the domestic prevent continued use of such tactics, the effectiveness of extended markets of their sponsors is permitted to private Western entities; deterrence of hybrid warfare is mixed at best and cannot be said to instituting an international fund that provides short-term financial be high. relief to nations suffering losses as a result of acts of economic pun- Surprise: Achieving tactical surprise is one of the principal ishment; and mounting continuing international legal challenges to reasons that hybrid warfare tactics are employed in the first place. attempted territorial rearrangements. Because the use of hybrid warfare tactics has largely remained Paramilitary hybrid warfare tactics are often used in an unchallenged and because U.S. adversaries enjoy shorter lines of attempt to obscure the identity of the instigator of a conflict. The communication and can be masters of strategic deception,60 the uniforms worn by insurgents, the weapons issued to them, social probability of further future tactical surprise due to the use of media postings, geolocation of the mobile phones of “volunteers” hybrid warfare tactics must be judged to remain high. sent into a conflict zone by the instigating state, and photographs Technology: In traditional land warfare, the ability to concen- of military equipment given to insurgents or used in support of trate forces is important to achieving tactical victory. The attacker- insurgents can and have all been used to put the lie to the meme to-target ratio is therefore high in hybrid warfare. Because hybrid that these individuals are acting outside of the control of the insti- warfare forces are organized ad hoc, the defender with a larger gator. Software might be developed to expose social media trolls formal army has higher reserves or strategic slack. In the hands of from the cyber domain that are being employed for the purposes a competent defender, both of these factors should militate in favor of IO in support of hybrid warfare operations in real time. Dis- of successfully deterring hybrid warfare. crediting opponent troll armies with their unwitting audiences is a Doctrine: The lack of a clearly articulated and salient doctrine form of cross-domain deterrence by denial. Propaganda campaigns by which to counter hybrid warfare tactics means that—despite and trolling are designed to manipulate intervention thresholds by President Obama’s two successful Scarborough Shoal and Senkaku sowing dissension in domestic political ranks. They are intended to Islands interventions—both the United States’ credibility and repu- complicate a deterrer’s ability to mobilize public opinion in support of a timely response to the aggressor. The United States’ and allies’

13 Cold War capabilities built to provide early warning and expose the aggressor that the deterrer is willing and able to react rapidly, such efforts might be revived. By providing repeated explanation thereby denying the aggressor the advantage of tactical surprise.62 of the tactics that adversaries are using and by providing warnings Such a deployment denies the aggressor its goal of manipulating or about those tactics’ potential repeated future use, Western govern- compromising the deterrer’s intervention and escalation thresholds. ments might “inoculate” Western audiences against the future, By making it clear that the deterrer is able to escalate rapidly to the repeated use of such tactics, thereby reducing their countries’ point of a full-blown military response, such a move also achieves vulnerability to them. If carried out on a sufficient scope and scale, immediate deterrence through the credible threat of rapid future this activity would likely promote the development of “herd immu- punishment (in-domain punishment). The extended deterrent value nity”61 in Western audiences. Both of these measures are in-domain of such forces can be enhanced by concluding advance agreements strategies of deterrence by denial. with nations neighboring potential trouble spots to host QRFs Hybrid warfare tactics also attempt to create ambiguity on during times of crisis. In-domain deterrence of the use of eco- the ground to make a strong response on the part of the deterrer nomic hybrid warfare tactics by means of the threat of punishment look disproportionate in the eyes of the international community. could entail accelerated punitive processes within the framework This strategy might be countered by means of a measured initial of the World Trade Organization (WTO) to counter trade-related response. Heavily armed police would initially be sent into a hybrid economic coercion and an international convention proscribing warfare conflict zone to augment local law enforcement, thereby acts of economic coercion and punishment and providing for a set providing a proportionate initial response to the outbreak of con- of predefined but flexible collective retaliatory measures against flict. These police units would be supported by military quick reac- aggressors (with the Gordian challenge here being adequately to tion forces (QRFs) temporarily deployed to neighboring countries define when such retaliatory measures are first triggered). in the region. The ability to conduct a “show of force” by airlifting Cross-domain deterrence: Manicheism, in which one sees a QRF into an adjacent country immediately upon the outbreak things as either good or evil, is a rather blunt diplomatic instru- of a crisis involving the use of hybrid warfare tactics signals to ment by which to counter threshold manipulation. This approach is best summarized by President George W. Bush’s message to foreign nations after the September 11, 2001, attacks: “Either you are The greater a regional power’s assertiveness, with us, or you are with the terrorists.” Because there is no room the greater the number of neighboring in Manicheism for shades of gray, it would deny the opponent the countries that will be looking to balance it by opportunity to operate in gray zones. The doctrine of “culpable negligence” is a more nuanced version of Manicheism. Under this means of closer alignment with an external doctrine, a state is deterred from allowing its citizens to volunteer great power. to destabilize a neighboring country by the cross-domain threat of

14 the punishment of being held to account internationally for refus- that of an opponent that has valuables that can be held at risk— ing to exercise adequate control over its population.63 The greater a applies at best only indirectly. regional power’s assertiveness, the greater the number of neighbor- Types of deterrence: In the context of terrorism, general ing countries that will be looking to balance it by means of closer deterrence can be said to be in force when such tactics are not alignment with an external great power. U.S. and allied diplomacy being employed to challenge the balance of power or mount attacks might be postured in such a way as to take maximum advantage on others. Immediate deterrence means the successful employ- of any such opportunities as they present themselves—a form of ment of measures to prevent the further employment of terrorist cross-domain deterrence by the threat of punishment. Software tactics after their first use. In view of the continuous and ongoing designed to identify and expose opposition troll armies also offers international military campaigns against the Islamic State in Iraq the prospect of achieving cross-domain deterrence by punishment and Syria, al Qaeda, al Mourabitoun, Boko Haram, al Shabaab, by means of the threat of nonkinetic cyberattacks on such actors. Abu Sayyaf, the Islamic Movement of Uzbekistan, and the Taliban, among others, in the Levant; in Syria and Iraq; in North, East, and Nonstate Actors West Africa; and in South and East Asia, the effectiveness of U.S. The individuals or organizations included in the termnonstate and allied general and immediate deterrence against nonstate actors actors can include virtual networks (such as hackers), nongovern- employing terrorist tactics can be said to be low. Direct deterrence mental organizations, civil society organizations, criminal groups of terrorism entails preventing attacks on the U.S. homeland. No or cartels, terrorist organizations, multinational corporations, major, mass-casualty attacks on the United States have recurred regional supranational economic organizations (such as the Euro- since 9/11. However, a significant number of events, such as those pean Union), and international organizations (such as the United at Fort Hood in November 2009, Boston in April 2013, San Nations).64 This section focuses on deterring the persistent threat Bernardino in December 2015, and New York City in November posed by (transnational) groups of nonstate actors that employ 2017, collectively involving over 370 casualties, have taken place. asymmetric terrorist tactics of warfare.65 While a far greater number of attacks might have taken place had Enablers: When applied to deterring nonstate actors employ- the United States not strengthened homeland security after the ing terrorist tactics, a major portion of the assumptions and 9/11 attacks, direct deterrence of attacks on the U.S. homeland still enabling conditions for successful deterrence appear not to be met cannot be said to be high. Extended deterrence of terrorism entails or seem to be absent. Because certain nonstate actors are will- preventing terrorist attacks on allies through the threat of retali- ing to sacrifice their lives in pursuit of their cause by committing ation. Acts of aggression by nonstate actors using terrorist tactics suicide (even if this is not generally a particularly effective method against a significant number of U.S. allies have recurred with of attack), one of the principal assumptions of deterrence theory— some regularity since the 9/11 attacks. Such attacks also continue to be mounted despite the fact that the United States is currently

15 engaged against their instigators militarily in multiple theaters of in employing asymmetric terrorist tactics. Deepened intelligence military operations. Because U.S. intervention on behalf of friends cooperation, random searches in public places, and periodic and allies against terrorist organizations in multiple theaters of random surges in the level of security at obvious targets increase operations has not attenuated the use of terrorist tactics of warfare, the probability that attackers will be thwarted, thereby reducing the U.S. capacity for extended deterrence of aggression by nonstate the risk of tactical surprise. Precluding the possibility of positive actors using terrorist tactics can be said to be low.66 publicity and ensuring negative media coverage instead could Surprise: In the context of terrorism, tactical surprise involves reduce the expected value to the aggressor of mounting an attack. A receiving warning of an attack but failing to have time to take concerted, international strategic communications campaign could measures to move potential victims out of harm’s way or forestall raise the expected costs and lower the expected benefits of attacks the attack. Open, Western societies are replete with potential soft by nonstate actors by emphasizing the following: targets susceptible to terrorist attack. Short of draconian repressive 1. the Islamic illegitimacy of such tactics, when Islam is abused measures or the prohibitively expensive “hardening”67 of potential to justify them targets throughout entire societies, further tactical surprise at the 2. the low success rate of such attacks hands of nonstate actors employing terrorist tactics appears to be 3. the failure of such campaigns to achieve their political inevitable. objectives Technology: The incentive for nonstate actors to strike first 4. their counterproductive nature, stigmatizing Muslims and using terrorist tactics is high. Repeated attacks have shown that causing sanctions the attacker-to-target ratio is low: It only takes one bomber to kill 5. empirical evidence that such attacks usher hardline politi- dozens; three terrorists killed 90 civilians during a November 2015 cians, less inclined to compromise, into office.69 attack at the Bataclan theater in Paris. Furthermore, the components needed to construct improvised explosive devices and individ- While collective responsibility and collective punishment are uals willing to assemble, deliver, and detonate them continue to be widely thought to be a cultural taboo in the West, cultural anthro- available in abundance. Strategic slack thus favors the attacker. pologists will attest that the same is not true of other cultures.70 As Doctrine: A well-articulated and broadly disseminated U.S. Boaz Ganor points out, measures taken against those who knew national doctrine by which to deter attacks by nonstate actors does about and did not prevent an attack and those who participated in not appear to exist. The salience and credibility of U.S. doctrine preparations and planning cannot be regarded as collective punish- are, therefore, low, as is the United States’ reputation for deterrence ment.71 Historically, families of nonstate actors that have executed in the nonstate actor subarea of the land domain of warfare.68 terrorist attacks have been granted pensions, compensation, and In-domain deterrence: In-domain strategies of deterrence jobs while the perpetrators have been celebrated in propaganda can be used to deny nonstate actors the advantages that they seek produced by supporting organizations that are funded in part by

16 the U.S. government. In-domain deterrence by punishment might Because they involve profound moral tensions and dilemmas, start by interdicting the provision of pensions, compensation, and the counterterror options described need to be thoroughly under- jobs as rewards to the families of nonstate actors who have commit- stood and analyzed not just through pragmatic lenses but through ted atrocities.72 Punishment could also extend to travel bans on the moral and ethical lenses as well. While the options mentioned do family members of both the perpetrators of acts of terror and the constitute theoretical possibilities by which to deter acts of terror, enablers. Clearly, holding the families of perpetrators and enablers the fact that they are mentioned here does not constitute a recom- collectively responsible for acts of terror raises serious questions. In mendation that they should actually be employed unless and until a hard-nosed example, Israel once regularly razed the homes of ter- significant and thorough further ethical examination and debate of rorists. This policy was the subject of debate within Israel in terms their effects and implications for U.S. international moral standing of its morality and of its actual effectiveness in deterring acts of has taken place. terror.73 In theoretical terms, however, family members are one of the few objects of value to perpetrators of acts of terror that might Cyberspace75 be held at risk in order to achieve better immediate deterrence.74 Enablers: The assumptions and requirements for successful Cross-domain deterrence: Nonstate actor groups frequently deterrence mostly appear to be met in the cyber domain. Some organize into networks of varying types of different cells and links opponents do, however, appear to have a relatively high appetite for between such cells. Such networks could be mapped, identify- taking risk in the cyber domain. This implies that it will, at least ing and monitoring important network bases and courier links. initially, be more difficult to deter such actors from future acts of Members of these networks might be deterred by the threat of cyber aggression. It is difficult to determine whether opponents’ cross-domain punishment by nonkinetic and kinetic means. In behavior in the cyber domain can be explained as recklessness born first order, the network of financiers that provide the funding that out of insufficient experience with the limitations and side effects nonstate actor organizations require to continue to function can be of such warfare or whether it is the result of cold, thorough calcula- identified and interdicted. The couriers that deliver such finances tion by opponents that they would be advantaged by escalation in from safe rear areas, such as the Persian Gulf, to the front lines in this domain.76 the Middle East or North Africa can be interdicted as well. The Types of deterrence: Because attempts to change the balance network of radicalizers that incites and recruits foot soldiers willing of power in the cyber domain have been under way for some time, to commit terrorist acts can be mapped and interdicted in much general deterrence can be considered to be quite low. Due to the the same fashion. Originating as it does in the domain of cyber- fact that we have witnessed repeated and continuing instances of space, the kinetic type of cross-domain deterrence by punishment both opponent computer network exploitation (spying and stealing often relies on another domain for its ultimate execution: drone of information) and opponent computer network attack,77 immedi- strikes from the air. ate deterrence of cyberwarfare can also be deemed to be low.

17 Surprise: With the possible exception of zero-day exploits (the mining which type of attack is likely to be carried out by which exploitation of previously unknown computer operating system type of attacker. Types of attack include IO by state-sponsored or software weaknesses), the probability of strategic surprise in entities; doxxing, whereby an individual’s personal information is the cyber domain looks low; the threat is well-known.78 Barring a deliberately made public to embarrass or endanger that individual; disarming first strike on both the commercial and governmental web-based confidence tricks; the theft of personally identifiable cyber defense resources of the United States or its allies, the ability information (PII) or intellectual property rights (IPR); heists, such to mobilize resources in response to a cyberattack appears to be as the theft of multimillion-dollar amounts from banks and central high, and the probability of tactical surprise therefore appears to be banks; the implantation of malware payloads on target computers; low. However, the incentives and resources required for the private the theft of government secrets; and computer network attack. The sector to protect critical infrastructure against cyberattack are sub- types of attacker include hobbyists; “hacktivists” (political activists stantial and likely not in place.79 who are active on the Internet); petty criminals stealing hundreds Technology: The United States’ ability to deter opponents or thousands of dollars by means of cyberattack; “great train rob- directly within the cyber domain is a function both of the preva- bers” pulling off multimillion-dollar heists of the kind described lence of networked computers in the target country and of that above; private-sector proxies hired by governments for purposes of country’s degree of interconnectedness with the outside world. For deniability, such as criminal gangs; terrorists; intelligence agencies; some countries (e.g., Russia, China), therefore, opportunities for and military units, such as People’s Liberation Army Unit 61398. direct deterrence may be high; for others (e.g., North Korea), they Once again, the black lines in Figure 4 are notional escalation may be lower. thresholds below or to the right of which the attacker or the type of Doctrine: The problem with intervention thresholds is not attack may be serious enough to warrant a military response. The so much that they can be manipulated or compromised but is bottom-right quadrant of this notional cyber threat matrix suggests instead an issue of attribution—the ability or willingness of the that the most threatening actors are military units, intelligence United States and its allies definitively to identify the ultimate agencies, and state proxies. Figure 4 suggests that the compromise actor that chose to cross an intervention threshold is limited. While of PII, the theft of IPR, and the implantation of malware payloads the United States has articulated a cyber deterrence doctrine, for present the greatest threats. the reasons given above, its credibility and reputation in deterring In-domain deterrence: At the network level, a number of opponent activity are low.80 measures are available. Much as telephone companies deliberately As in the case with space attacks, cyberattacks can be disag- instruct telephone exchanges not to permit incoming calls to areas gregated in terms of the level of threat posed by the actors carrying hit by natural disasters to prevent network overload, a collective them out and the types of attacks that those actors might execute international legal mechanism could be created to deny inter- (Figure 4). The greatest potential threats can be identified by deter- national Internet backbone access to conflict parties in times of

18 Figure 4. Cyberspace Threat Matrix

Seriousness of intrusion, exploitation, or attack

Percentage

IO Doxxing Cons PII IPR Heists Malware Secrets Attack with Hobbyists 66% Hacktivists 66% Petty criminals 33% Great train robbers 33%

Resources of actor Proxies 77% Terrorists 33% Intelligence 88% Military 88% Percentage with 63% 50% 63% 75% 75% 50% 75% 63% 33%

NOTES: The check marks indicate whether the attackers identified in each row can carry out the type of disruption or attack indicated in each column. The percentage of attackers that can carry out each type of attack is then calculated for each column (x-axis percentages). Similarly, the percentage of types of attack that each type of attacker can carry out is calculated for each row (y-axis percentages). crisis. While fraught with challenges concerning the conditions principles),83 no-notice red team attacks, and regular continuity under which it might fi rst be triggered, such an instrument might of service exercises might improve organizational cyber robustness improve immediate deterrence in the cyber domain; most forms and cyber resilience. of cyberwarfare require Internet access.81 Other preventive steps At the level of individual computing devices, in-domain include actively shaping network topology to reduce the number of deterrence by denial could continue by working with the insur- high-degree nodes and ensuring that all data that have to be stored ance industry to promulgate a national device robustness standard in a network-accessible fashion are encrypted.82 Deterrence by implemented by national testing laboratories, such as Underwriters denial continues at the governmental and organizational levels with Laboratories.84 Th e United States and allied governments could the use of extremely robust, highly connected server clusters that use their monopsony85 market power to promulgate fi rmware migrate between various previously unknown network (“darknet”) solutions on top of low-security, legacy Internet communications clouds (“dark clouds”) and provide seamless, emergency continuity protocols that ensure a very high level of confi dence in the identity of web and other computer services. Annual organizational cyber of the user on the other end of a computer connection and reliable audits (a possible future Financial Accounting Standards Board encryption when handling high volumes of sensitive (government) requirement for a clean audit under generally accepted accounting data.

19 Conclusion Because they do not pose a risk of escalating While it has not addressed the challenge of containing horizon- conflict and because they can be reversed, tal escalation risks, this Perspective has described a large number thereby permitting an exit from the conflict of possible in-domain and cross-domain approaches by which to to the status quo ante, nonescalatory, implement a doctrine that might be able to contain vertical and lateral escalation risks across three domains of military activity. reversible strategies of deterrence are most Approximately one-third of the suggested strategies are cross- preferred. domain strategies. How might one prioritize this long list of sug- gestions? Because the enablers of successful deterrence identified by classical texts are least present in these two areas, establishing At the level of the individual user, biometric certificates replac- effective deterrence of attacks in space and of the use of hybrid ing Social Security numbers can reduce cyber fraud significantly, warfare tactics are the most urgent priorities. Measures by which to and regular training and recertification can ensure better cyber rectify significant vulnerabilities in the space domain include the hygiene and lower susceptibility to social engineering attacks following: (attacks that trick computer users into revealing their passwords • achieving bipartisan, executive-legislative consensus to put or other critical PII).86 Martin Libicki has explained in detail why policies in place that ensure that movement toward a more problems of attribution, unintended effects, and the difficulty of connected, distributed, robust, and resilient satellite network in battle damage assessment make deterrence by in-domain punish- space will take place over the long term ment a problematic proposition in the cyber domain.87 • demonstrating to opponents, by means of frequent allied Cross-domain deterrence: A number of cross-domain mea- military exercises, an increasing ability to operate despite the sures by which to deter cyberattack using the threat of punishment degradation of space-based assets do exist, however. Because certain opponents claim jurisdiction • concluding an agreement on detailed criteria that would trigger over all communications that enter or leave their country through a collective response against attacks on the space-based assets their international gateways, they can be referred to the WTO for of the United States and parties allied with the United States the piracy of IPR in violation of the WTO Trade-Related Aspects through their signature of a treaty for the collective defense of of Intellectual Property Rights agreement.88 Collective, interna- assets in space tional cyberdefense agreements are another method of punishment • taking visible steps to map and hold at risk the infrastructure available both to protect government networks, systems, and data by which adversaries create a protected national information and to protect private-sector IPR.89 space—identifying the organizations, computer systems, and

20 other equipment that filter or block web content and visibly (nonescalatory deterrent strategies) are preferable to those that preparing to attack them would cause escalation.90 • strengthening the U.S. and allied ability to reach mass audi- • Prefer reversible to irreversible measures—Because deter- ences within adversary information spaces rence is costly and the ability to de-escalate a conflict con- • making it clear that opponent information control, C3ISR, sciously is important to successful crisis management, deter- and RSTA infrastructure will suffer significant damage—in rent strategies that are reversible are, as a general proposition, other words, coming up with credible threats of punishment. preferable to those that are not. • Prefer denial to punishment—We know that, as a general Urgent measures that can be taken to deter further use of proposition, deterrence by denial is to be preferred to deterrence hybrid warfare include the following: by punishment because the latter requires continuous coercion, • an enhanced ability to identify and interdict troll armies whereas the former involves control.91 • an enhanced ability to inoculate the public against IO • greater efforts at speedy attribution of the origin of combatants In a first step, the “nonescalation” and “reversibility” just men- • agreement on detailed criteria that would trigger a rapid allied tioned can be combined to provide an ordinal ranking of deterrent response strategies (Figure 5). Because they do not pose a risk of escalating • a visible and credible capability to deploy both heavily armed conflict and because they can be reversed, thereby permitting an police and supporting military QRFs rapidly to crisis areas and exit from the conflict to thestatus quo ante, nonescalatory, revers- neighboring states ible strategies of deterrence are most preferred. Nonescalatory but • advance agreement with neighboring states to host military nonreversible strategies come next. Escalatory but reversible strate- QRFs that might be moved to support police forces in crisis gies follow, and nonreversible escalatory strategies take up the rear areas, if needed. of the pack.

Beyond these urgent measures, the examination of classical Figure 5. Ordinal Ranking of Deterrent Strategies deterrence theory offers three broad filtering criteria that might be applied roughly to prioritize the remaining strategies that have been Reversible Nonreversible suggested: • Prefer nonescalatory to escalatory approaches—Generally, Nonescalatory 1 2 but not always, strategies that offer the prospect of respond- Escalatory 3 4 ing to or deterring an opponent without escalating the conflict

21 In a second step, the criterion according to which strategies Even when limited to Russia, China, and counterterrorism, as of denial are preferred over those of punishment can be added to suggested, achieving effective cross-domain deterrence has sig- provide a finer ranking (Figure 6).92 The strategies most preferred nificant organizational, diplomatic, and resource implications. A are those involving deterrence by denial that are nonescalatory and review of the strategies identified shows that almost half of them reversible. Denial using nonescalatory, nonreversible strategies fol- rely for their execution on nonmilitary organizations. The civilian lows. Punishment using nonescalatory, reversible strategies comes organizations involved include domestic and foreign civilian intel- next. Punishment employing nonescalatory but nonreversible ligence agencies, the U.S. Department of State and foreign minis- strategies follows. Only once the nonescalatory options have been tries of foreign affairs, the U.S. Broadcasting Board of Governors exhausted do we turn to escalation: first seeking to deter by denial and its foreign counterparts, U.S. and international Tier 1 Internet using reversible, escalatory strategies, then using nonreversible, backbone providers, the U.S. Financial Accounting Standards escalatory strategies. Deterrence by the threat of punishment using Board and its international counterparts, the insurance industry, reversible escalatory strategies is among the last resorts. Irrevers- and national testing laboratories. ible, escalatory punishment is the least preferable option. For the A smaller, but significant, number of the strategies discussed reasons given above, this ranking of strategies should be regarded involve collective action in concert with other friendly or allied as a rough guide, not a hard and fast rule. Context, timing, and nations. They include those involving strategic communications, opponent mindsets are important and can quickly scramble any diplomatic balancing, implementing a doctrine of culpable negli- rigid dictates of doctrine. gence, and collective security agreements that defend against economic measures short of war, space attack, cyberattack, and hybrid warfare. These facts suggest the following: Figure 6. Partitioning the Set of Deterrent Strategies • Policymakers may need to spend political capital within both national and international stakeholder groups to build consen- Reversible Nonreversible sus on the need for action. This may entail elaborating and achieving consensus within and Denial across national and allied defense and diplomatic establishments on Nonescalatory Punishment a concept of operations by which to implement a doctrine of cross- domain deterrence. A significant subsequent international strategic Denial communications effort that popularizes and wins public support for Escalatory such a concept might have to follow. Punishment • Decisionmakers may also need to consider reallocating national human and financial resources in such a way as to ensure that

22 the entities being relied on to execute the strategy are properly resourced, highly interoperable, and very likely to achieve unity Irreversible, escalatory punishment is the of national (and international) effort. least preferable option. Entities, such as the U.S. Department of State, that will have to be relied on to reinvigorate existing and forge new collective • The United States may need to reallocate resources within security agreements may need to receive additional resources government departments to bolster bilateral relationships with at the expense of other government departments. Bureaucratic its allies. reorganizations that cross departmental boundaries and merge Significant efforts might need to be made with individual U.S. departmental functional and geographical offices into national allies to achieve international political consensus and to adapt alli- centers of competence should not be taboo. The overall value ances to the changed threat profile. If these efforts are to have any to the U.S. national interest, and the organization, person- prospect of success, high-quality human resources would need to be nel, and resourcing of existing supranational collective secu- dedicated to them. The resources required likely exceed those cur- rity bureaucracies, such as NATO, might also be scrutinized rently dedicated to bilateral, allied, and politico-military diplo- rigorously in light of the new demands of the changed strategic macy materially—both in qualitative and in quantitative terms. landscape. • To ensure their effectiveness, resources may need to be • Organizations focusing on strategic communication and reallocated toward international bodies that the United States collective defense negotiations that the United States dises- and its allies would rely on for the execution of significant tablished at the end of the Cold War may need to be rees- parts of these strategies. tablished out of existing resources in some streamlined and Such organizations as the United Nations, the International updated form. Atomic Energy Agency, the Organization for the Prevention of For the United States, this might entail the reestablishment of cog- Chemical Warfare, and other, new supranational bodies puta- nates of the U.S. Information Agency and the Arms Control and tively created to deter cyberwarfare, war in space, and economic Disarmament Agency. Tailored to meet the significantly changed warfare and to achieve more-effective Western strategic communi- demands of modernity, any such entities would likely bear little cation may need to have their existing funding focused on priority resemblance to their predecessors. Nor is it self-evident from the areas that would be relied on to implement such a doctrine or may outset that the Department of State would be the natural home for need to receive new funds. such entities.

23 References Davis II, J. S., M. C. Libicki, S. E. Johnson, J. Kumar, M. Watson, and A. Karode, A Framework for Programming and Budgeting for Cybersecurity, Santa Ablodia, T., “Shevardnadze Says Moscow Backs Rebels,” The Independent, Monica, Calif.: RAND Corporation, TL-186-DHS, 2016. As of January 31, 2018: March 17, 1993. As of January 9, 2018: https://www.rand.org/pubs/tools/TL186.html http://www.independent.co.uk/news/world/europe/ shevardnadze-says-moscow-backs-rebels-1498157.html Davis, P. K., Some Lessons Learned from Building Red Agents in the RAND Strategy Assessment System (RSAS), Santa Monica, Calif.: RAND Corporation, Abrams, M., “Deterring Terrorism: A New Strategy,” Perspectives on Terrorism, N-3003-OSD, 1989. As of January 31, 2018: Vol. 8, No. 3, 2014, p. 13. https://www.rand.org/pubs/notes/N3003.html Air Force Studies Board, U.S. Air Force Strategic Deterrence Analytic Capabilities: Davis, P. K., Toward Theory for Dissuasion (or Deterrence) by Denial: Using Simple An Assessment of Tools, Methods and Approaches for the 21st Century Security Cognitive Models of the Adversary to Inform Strategy, Santa Monica, Calif.: RAND Environment, Washington D.C.: The National Academies Press, 2014. Corporation, WR-1027, 2014. As of January 31, 2018: Asmus, R. D., A Little War That Shook the World: Georgia, Russia, and the Future https://www.rand.org/pubs/working_papers/WR1027.html of the West, 1st ed., New York: Palgrave Macmillan, 2010. Davis, P. K., and B. M. Jenkins, Deterrence and Influence in Counterterrorism: A Blackwill, R. D., and J. M. Harris, War by Other Means: Geoeconomics and Component in the War on al Qaeda, Santa Monica, Calif.: RAND Corporation, Statecraft, Cambridge, Mass.: The Belknap Press of Harvard University Press, MR-1619-DARPA, 2002. As of January 31, 2018: 2016. https://www.rand.org/pubs/monograph_reports/MR1619.html Bond, D., and A. Wisniewska, “Terror Attacks on Developed Nations Hit 16-Year Defense Science Board, Task Force on Cyber Deterrence, Washington, D.C.: U.S. High,” Financial Times, November 14, 2017. As of January 31, 2018: Department of Defense, 2017. https://www.ft.com/content/3c258898-c95c-11e7-ab18-7a9fb7d6163e Delpech, T., Nuclear Deterrence in the 21st Century: Lessons from the Cold War Brodie, B., Strategy in the Missile Age, Princeton, N.J.: Princeton University Press, for a New Era of Strategic Piracy, Santa Monica, Calif.: RAND Corporation, 1959. MG-1103-RC, 2012. As of January 31, 2018: https://www.rand.org/pubs/monographs/MG1103.html Clarke, R. A., and R. K. Knake, Cyber War: The Next Threat to National Security and What to Do About It, 1st ed., New York: Ecco, 2010. DoD—See U.S. Department of Defense. Cohen, A. E., and R. E. Hamilton, The Russian Military and the Georgia War: Dou, E., “China’s Great Firewall Gets Taller,” Wall Street Journal, January 30, Lessons and Implications, Carlisle, Pa.: Strategic Studies Institute, U.S. Army War 2015. As of January 31, 2018: College, 2011. https://www.wsj.com/articles/chinas-great-firewall-gets-taller-1422607143 Connable, B., J. H. Campbell, and D. Madden, Stretching and Exploiting Edelman, E. S., and H. Brands, Why Is the World So Unsettled? The End of the Thresholds for High-Order War: How Russia, China, and Iran Are Eroding American Post-Cold War Era and the Crisis of Global Order, Washington, D.C.: Center for Influence Using Time-Tested Measures Short of War, Santa Monica, Calif.: RAND Strategic and Budgetary Assessments, 2017. As of January 31, 2018: Corporation, RR-1003-A, 2016. As of January 31, 2018: http://csbaonline.org/uploads/documents/ https://www.rand.org/pubs/research_reports/RR1003.html Why_Is_the_World_So_Unsettled_FORMAT_FINAL.pdf Cooper, Z., and J. Douglas, “Successful Signaling at Scarborough Shoal?” War on Eilperin, J., “Obama: U.S. Stands by Treaty with Japan, but Diplomacy Is Way the Rocks, May 2, 2016. As of January 31, 2018: to Settle Dispute over Islands,” Washington Post, April 24, 2014. As of January 31, https://warontherocks.com/2016/05/successful-signaling-at-scarborough-shoal/ 2018: https://www.washingtonpost.com/world/ Cornell, S. E., and S. F. Starr, The Guns of August 2008: Russia’s War in Georgia, president-obama-affirms-us-will-stand-by-treaty-obligations-to-japan/ Armonk, N.Y.: M.E. Sharpe, 2009. 2014/04/24/425dd9c8-cb62-11e3-93eb-6c0037dde2ad_story.html

24 Farwell, J. P., and D. Arkelian, “China Cyber Charges: Take Beijing to the WTO Libicki, M. C., Cyberdeterrence and Cyberwar, Santa Monica, Calif.: RAND Instead,” The National Interest, May 20, 2014. As of January 31, 2018: Corporation, MG-877-AF, 2009. As of January 31, 2018: http://nationalinterest.org/blog/the-buzz/ https://www.rand.org/pubs/monographs/MG877.html china-cyber-charges-take-beijing-the-wto-instead-10496 Lindsay, J. R., and E. Gartzke, “Cross-Domain Deterrence as a Practical Problem Feaver, P., “What Is Grand Strategy and Why Do We Need It?” Foreign Policy, and a Theoretical Concept,” in J. R. Lindsay and E. Gartzke, eds.,Cross-Domain April 8, 2009. As of January 31, 2018: Deterrence: Strategy in an Era of Complexity, San Diego, Calif.: University of http://foreignpolicy.com/2009/04/08/ California, San Diego, 2016. what-is-grand-strategy-and-why-do-we-need-it/ Malina, B. J., The New Testament World: Insights from Cultural Anthropology, Freedman, L., The Evolution of Nuclear Strategy, 3rd ed., Basingstoke, UK: Louisville, Ky.: Westminster John Knox Press, 2001. Palgrave Macmillan, 1981. Manzo, V., “Deterrence and Escalation in Cross-Domain Operations: Where Freedman, L., Deterrence, Cambridge, UK: Polity Press, 2004. Do Space and Cyberspace Fit?” Strategic Forum, No. 272, Washington, D.C.: National Defense University, Institute for National Strategic Studies, Ganor, B., The Counter-Terrorism Puzzle: A Guide for Decision Makers, Herzliya, December 2011. Israel: The Interdisciplinary Center, International Policy Institute for Counter- Terrorism, 2005. McGuire, W. J., “Resistance to Persuasion Conferred by Active and Passive Prior Refutation of Same and Alternative Counterarguments,” Journal of Abnormal George, A. L., and R. Smoke, Deterrence in American Foreign Policy, New York: Psychology, Vol. 63, No. 2, 1961. Columbia University Press, 1974. Miller, S. E., and S. Van Evera, Naval Strategy and National Security: An Gorman, S., A. Cole, and Y. Dreazen, “Computer Spies Breach Fighter-Jet “International Security” Reader, Princeton, N.J.: Princeton University Press, 2014. Project,” Wall Street Journal, April 21, 2009. As of January 31, 2018: https://www.wsj.com/articles/SB124027491029837401 Ministry of Defense of the Russian Federation, “Ʉɨɧɰɟɩɬɭɚɥɶɧɵɟ ɜɡɝɥɹɞɵɧɚɞɟɹɬɟɥɶɧɨɫɬɶɜɨɨɪɭɠɟɧɧɵɯɫɢɥɊɨɫɫɢɣɫɤɨɣɎɟɞɟɪɚɰɢɢɜ Hitchens, T., and J. Johnson-Freese, Toward a New National Security Space ɢɪɮɨɪɦɚɰɢɨɧɧɨɦɩɪɨɫɬɪɚɧɫɬɜɟ” [“Conceptual Views on the Activities of the Strategy: Time for a Strategic Rebalancing, Washington, D.C.: Atlantic Council, Armed Forces of the Russian Federation in the Information Space”], 2011. As of 2016. As of January 31, 2018: January 31, 2018: http://www.atlanticcouncil.org/images/publications/ http://ens.mil.ru/science/publications/more.htm?id=10845074%40cmsArticle AC_StrategyPapers_No5_Space_WEB1.pdf Morgan, F. E., Deterrence and First-Strike Stability in Space: A Preliminary Huth, P. K., Extended Deterrence and the Prevention of War, New Haven, Conn.: Assessment, Santa Monica, Calif.: RAND Corporation, MG-916-AF, 2010. As of Yale University Press, 1991. January 31, 2018: Iklé, F. C., “The Reagan Defense Program: A Focus on the Strategic Imperatives,” https://www.rand.org/pubs/monographs/MG916.html Strategic Review, Vol. 10, No. 2, 1982. Morgan, P. M., Deterrence Now, Cambridge, UK: Cambridge University Press, Kahn, H., On Escalation: Metaphors and Scenarios, New York: Praeger, 1965. 2003. Kahneman, D., and A. Tversky, “Prospect Theory: An Analysis of National Institute of Standards and Technology, Framework for Improving Critical Decisionmaking Under Risk,” Econometrica, Vol. 47, No. 2, 1979. Infrastructure Cybersecurity, Version 1.0, Washington, D.C., February 12, 2014. As of January 31, 2018: Kroenig, M., and B. Pavel, “How to Deter Terrorism,” The Washington Quarterly, https://www.nist.gov/sites/default/files/documents/cyberframework/ Vol. 15, Spring 2012. cybersecurity-framework-021214.pdf Lehman, J. F., “Rebirth of a U.S. Naval Strategy,” Strategic Review, Vol. 9, No. 3, 1981.

25 NATO Allied Command Transformation, “NATO Countering the Hybrid Posen, B., Restraint: A New Foundation for U.S. Grand Strategy, Ithaca, N.Y.: Threat,” 2011. As of January 31, 2018: Cornell University Press, 2014. http://www.act.nato.int/nato-countering-the-hybrid-threat Quinlivan, J. T., and O. Oliker, Nuclear Deterrence in Europe: Russian Approaches Newman, M. E. J., Networks: An Introduction, Oxford, UK: Oxford University to a New Environment and Implications for the United States, Santa Monica, Calif.: Press, 2010. RAND Corporation, MG-1075-AF, 2011. As of January 31, 2018: https://www.rand.org/pubs/monographs/MG1075.html Nye, J., and S. Joseph, “Deterrence and Dissuasion in Cyberspace,” International Security, Vol. 41, No. 3, 2017. Rasmusen, E., Games and Information: An Introduction to Game Theory, 4th ed., Malden, Mass.: Blackwell Publishing, 2007. Obama, B., National Security Strategy, Washington, D.C.: Executive Office of the President, 2015. As of January 31, 2018: Ravich, S. F., “State-Sponored Cyberspace Threats: Recent Incidents and U.S. https://obamawhitehouse.archives.gov/sites/default/files/docs/ Policy Response,” testimony before the Senate Foreign Relations Committee, 2015_national_security_strategy_2.pdf Subcommittee on East Asia, the Pacific and International Cybersecurity, Washington, D.C., June 13, 2017. Olavsrud, T., “9 Best Defenses Against Social Engineering Attacks,” October 19, 2010. As of January 31, 2018: Roberts, B., The Case for Nuclear Weapons in the 21st Century, Stanford, Calif.: https://www.esecurityplanet.com/views/article.php/3908881/ Stanford University Press, 2016. 9-Best-Defenses-Against-Social-Engineering-Attacks.htm Rogozin, D. O. E., “Voina i mir v terminakh in opredeleniyakh: Voenno- Perlez, J., “Asian Leaders at Regional Meeting Fail to Resolve Disputes over South politicheskij slovar” [“War and Peace in Terms and Definitions: A Military- China Sea,” New York Times, July 12, 2012. As of January 31, 2018: Political Dictionary”], 2011. As of January 31, 2018: http://www.nytimes.com/2012/07/13/world/asia/ http://www.voina-i-mir.ru/article/249 asian-leaders-fail-to-resolve-disputes-on-south-china-sea-during-asean-summit. Rohrbaugh, R. E., The Social Sciences and New Testament Interpretation, Peabody, html Mass.: Hendrickson, 1996. Perlez, J., “China and Japan, in Sign of a Thaw, Agree to Disagree on a Disputed Schelling, T. C., Arms and Influence, New Haven, Conn.: Yale University Press, Island Group,” New York Times, November 7, 2014. As of January 31, 2018: 1966. https://www.nytimes.com/2014/11/08/world/asia/ china-japan-reach-accord-on-disputed-islands-senkaku-diaoyu.html Social Security Administration, Office of the Inspector General,The Social Security Administration Is Pursuing Matching Agreements with New York and Other Perlez, J., “In Victory for Philippines, Hague Court to Hear Dispute over South States Using Biometric Technologies, Washington, D.C., 2000. As of January 31, China Sea,” New York Times, October 30, 2015. As of January 31, 2018: 2018: https://www.nytimes.com/2015/10/31/world/asia/ https://www.scribd.com/document/1926804/Social-Security-9841007 south-china-sea-philippines-hague.html?_r=0 Scouras, J., E. Smyth, and T. Mahnken, Cross-Domain Deterrence in U.S.-China Perlez, J., “New Chinese Vessels Seen Near Disputed Reef in South China Sea,” Strategy, Laurel, Md.: Johns Hopkins Applied Physics Laboratory, 2014. As of New York Times, September 5, 2016. As of January 31, 2018: January 31, 2018: https://www.nytimes.com/2016/09/05/world/asia/ http://www.jhuapl.edu/ourwork/nsa/papers/CrossDomainWeb.pdf south-china-sea-philippines-scarborough-shoal.html Slayton, R., “What Is the Cyber Offense-Defense Balance? Conceptions, Causes, Pipes, R. A., and A. A. Zuehlke, Jr., “Correlation of Forces” in Soviet Usage—Its and Assessment,” International Security, Vol. 41, No. 3, 2017, pp. 72–109. Meaning and Implications, Arlington, Va.: Defense Advanced Research Projects Agency, 1978. As of January 31, 2018: Snyder, G. H., “Deterrence and Power,” Journal of Conflict Resolution, Vol. 4, http://www.dtic.mil/dtic/tr/fulltext/u2/a059430.pdf No. 2, 1960.

26 Snyder, G. H., Deterrence and Defense: Toward a Theory of National Security, Princeton, N.J.: Princeton University Press, 1961. Snyder, S., Negotiating on the Edge: North Korean Negotiating Behavior, Washington, D.C.: United States Institute of Peace Press, 1999. Tadelis, S., Game Theory: An Introduction, Princeton, N.J.: Princeton University Press, 2013. Trager, R. F., and D. P. Zagorcheva, “Deterring Terrorism: It Can Be Done,” International Security, Vol. 30, No. 3, 2005, pp. 87–123. Treverton, G. F., and S. G. Jones, Measuring National Power, Santa Monica, Calif.: RAND Corporation, CF-215, 2005. As of January 31, 2018: https://www.rand.org/pubs/conf_proceedings/CF215.html U.S. Department of Defense, Department of Defense Strategy for Operating in Cyberspace, Washington, D.C., 2011. U.S. Department of Defense, FM 3-38 Cyber Electromagnetic Activities, Washington D.C.: Department of the Army, 2014. U.S. Department of Defense, The DoD Cyberstrategy, Washington, D.C., April 2015. As of January 31, 2018: https://www.defense.gov/Portals/1/features/2015/0415_cyber-strategy/ Final_2015_DoD_CYBER_STRATEGY_for_web.pdf U.S. Department of Defense, DoD Dictionary of Military and Associated Terms, Washington, D.C., August 2017. As of January 31, 2018: http://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/dictionary.pdf Vaccines.gov, “Vaccines Protect Your Community,” 2017. As of January 31, 2018: https://www.vaccines.gov/basics/protection/index.html

27 Notes 12 Freedman, 2004, pp. 37, 39; Huth, 1991; G. H. Snyder, 1961, p. 14. A recent report published by the National Research Council (Air Force Studies Board, 1 DoD, 2017, p. 21. 2014, pp. 25, 34, 41) challenges Glenn Snyder’s classical dichotomy between 2 For a more detailed discussion of the breakdown of the post–Cold War order, see deterrence by punishment and deterrence by denial (from G. H. Snyder, 1960). Edelman and Brands (2017). The report argues that denial is much better seen as a form of dissuasion than as a form of deterrence. According to this argument, “denial” would belong under 3 For a formal definition of strategy sets, see Rasmusen (2007). Paul Davis points “induce” in Figure 1, instead of under “deter.” The report suggests that “broad out that the elements of the strategy spectrum depicted in Figure 1 resemble the deterrence” should be defined to include both “deterrence by punishment” and incentives and disincentives comprising “influence theory” introduced by George “dissuasion by denial.” See also P. K. Davis (2014). Under Laurence Freedman’s and Smoke (P. K. Davis, 2014, citing George and Smoke, 1974). taxonomy, however, “denial” might more appropriately be placed under “preven- 4 Freedman, 2004, p. 57; Huth, 1991, pp. 52, 65–67. tion,” as it is a form of control. 13 5 Freedman, 2004, pp. 57, 104. As Paul Davis points out, if “denial” is infeasible, unaffordable, or impossible without undercutting values, then deterrence by threat of punishment may be the 6 Freedman, 2004, pp. 26, 27. better option (P. K. Davis, 2014). 7 Freedman, 2004, pp. 85–86. Preventive war is “a premeditated attack by one 14 Freedman, 2004, p. 40; Huth, 1991, p. 22. country against another, which is unprovoked in the sense that it does not wait 15 upon a specific aggression or other overt action by the target state, and in which Freedman, 2004, p. 67. the chief and most immediate objective is the destruction of the latter’s overall 16 Huth, 1991, p. 16. military power and especially its strategic [forces]” (Brodie, 1959, p. 227). 17 For a definition of the termzero-sum game, see Tadelis (2013). 8 Preemption can only occur when the expected cost of an opponent’s first strike 18 exceeds the damage expected to follow from one’s own first strike. It requires a P. M. Morgan, 2003, pp. 42–78. substantial, if not decisive, first-strike advantage, thereby establishing an asymme- 19 As Paul Davis has pointed out, however, there are indeed decisionmakers who try in force that favors the attacker. The critical factor in preemption is the degree are cognitively biased toward taking risks. This is a function of character, tempera- of imminence of an opponent first strike perceived by the attacker. Varying assessment, and state. If they are desperate, or even in the “domain of losses” in psycho- ments of the imminence of an opponent first strike can negatively affect the politi- logical terms, they may be predictably prone to take greater risks than their own cal legitimacy and legality of preemption under international law. Opponents normal utility functions would justify (author communication with Paul Davis). thought impervious to deterrence, adversary activity indicative of preparation See also Air Force Studies Board (2014) and P. K. Davis (1989). for war, and a desire to forestall the escalation of a limited war can all be cause for preemption. If the adversary capabilities targeted survive a preemptive strike, 20 Empirical evidence shows that decisionmakers frequently violate the behavioral they will likely be retargeted to inflict maximum damage and used immediately axioms on which classical economic theory and the von Neumann-Morgenstern (Freedman, 2004, pp. 2, 4, 24, 87; Huth, 1991, p. 10; G. H. Snyder, 1961, pp. 60, expected utility theorem depend. Furthermore, stress and the shortness of time 104, 107, 108, 110). in crisis situations can lead to “cognitive rigidity” that degrades the quality of decisionmaking. Deterrence, at its least, presupposes an ability to maximize one’s 9 Freedman, 2004, pp. 26, 40, 100, 109, 110; Schelling, 1966, pp. 2, 4, 100, 172, own subjective expected utility based on a logical assessment of potential gains 174; G. H. Snyder, 1961, p. 40. and losses and an assessment of the probabilities of opponent actions. Some form 10 Freedman, 2004, p. 6; Huth, 1991, p. 15; Schelling, 1966, p. 13; G. H. Snyder, of continuing effort on each party’s part to search for new information and inte- 1961, p. 11. grate that information into calculations that assess the expected costs and benefits of alternative courses of action is also assumed (Air Force Studies Board, 2014, 11 Freedman, 2004, pp. 8, 61, 66; Schelling, 1966, pp. 183, 214; G. H. Snyder, p. 98; Freedman, 2004, pp. 22, 35–36, 49; Huth, 1991, pp. 30–31, 54, 137–138, 1961, pp. 15, 53, 76. 200, 203–204; Kahneman and Tversky, 1979; G. H. Snyder, 1961, pp. 10, 27; Trager and Zagorcheva, 2005). Davis discusses how relaxing these assumptions

28 when building “Red agents” for nuclear escalation simulation models can lead to 29 For the negative implications of a multipolar system for strategic stability, see more-realistic, richer, more-variegated, at times counterintuitive, and therefore Delpech (2012, p. 94) and Roberts (2016, p. 82). more-valuable analytical results (P. K. Davis, 1989, 2014). 30 For a discussion of the Soviet “correlation of forces” methodology of assessing 21 “Bargaining reputation can be defined as the perceived willingness of a state’s the strategic balance, see, for example, Pipes and Zuehlke (1978). political and military leadership to risk armed conflict in pursuit of foreign policy 31 objectives and the likelihood that it will accede to the demands of adversaries under “Grand strategy is a term of art . . . and refers to the collection of plans and policies that comprise the state’s deliberate effort to harness political, military, coercive pressure” (Huth, 1991, QBMTPTFF)VUI pp. 6, 9 BOE diplomatic, and economic tools together to advance that state’s national interest. 'SFFENBO QQ ). Grand strategy is the art of reconciling ends and means. It involves purposive 22 Strategic surprise occurs when a deterrer “fails to react to intelligence informa- action—what leaders think and want. Such action is constrained by factors leaders tion indicating a threat of attack and is militarily unprepared to counter the attack explicitly recognize (for instance, budget constraints and the limitations inherent when it is launched.” Tactical surprise occurs when a deterrer “receives warning of in the tools of statecraft) and by those they might only implicitly feel (cultural an attack but has insufficient time to mobilize fully and position forces for optimal or cognitive screens that shape worldviews)” (Feaver, 2009). Grand strategy is a defense against an impending attack” (Huth, 1991, p. 35, note 13). problematic concept. Some argue that it is a fictive strategy inferred post facto, not something that is decided upon ex ante. 23 Deterrence of an all-out nuclear attack on the deterrer or its protégés by the threat of severe punishment is likely to be perceived as “proportionate” and there- 32 Obama, 2015. fore to be thought to be both legitimate and quite credible. “But for lesser chal- 33 lenges, such as conventional attack on an ally, a threat to inflict nuclear punish- Doctrine is fundamental principles that guide the employment of United States ment [in response] may be perceived as disproportionate and indiscriminate and military forces in coordinated action toward a common objective and may include terms, tactics, techniques, and procedures (DoD, 2017). For a discussion of the therefore less credible than a threat to fight a ‘denial’ action” (Freedman, 2004, merits of a more “restrained” approach to U.S. grand strategy, see Posen (2014). pp. 33, 35, quoting G. H. Snyder, 1961, p. 15). 34 24 Crisis stability is “a measure of countries’ incentives not to preempt in a crisis, Russia’s choices of government and civilian targets among the 38 websites that it that is, not to attack first in order to beat the attack of the enemy” (F. E. Morgan, attacked during the 2008 Russo-Georgian war are instructive. The websites of the president, parliament, the foreign ministry, the interior ministry, and the national 2010, pp. 1–2, citing Charles Glaser). bank were targeted. So were the websites of private news agencies and banks. 25 Brodie, 1959, pp. 165, 177, 185; Schelling, 1966, p. 244. Unacceptable damage The local government and local news service websites of the town of Gori were is inflicted by means of a “counterforce” strike against adversary military units, by attacked in support of Russian tactical ground operations to take the town. Inter- a “counter-value” strike against opponent population centers, or by any number of net backbone links to Turkey and Ukraine on which Georgia is highly dependent associated measures on the “escalation ladder” located between these two extremes were subject to cyberattack, and so were the embassy websites of the United States (Kahn, 1965, p. 34). and the United Kingdom, lest they think of issuing statements in support of the Georgian government narrative of events in the parallel information war (Cohen 26 Kahn, 1965, p. 4. and Hamilton, 2011, p. 44). 27 Connable, Campbell, and Madden, 2016, pp. 19, 22; S. Snyder, 1999, p. 234. 35 Scholars disagree on a single, common definition ofcross-domain deterrence. 28 Connable, Campbell, and Madden, 2016, p. 19; S. Snyder, 1999, p. 234. Clar- DoD defines five domains: air, land, sea, cyber, and space (DoD, 2017). By con- ity of intervention and/or escalation thresholds is not a conditio sine qua non. In trast, some scholars define domains in terms of weapons and types of belligerents: certain circumstances, such as NATO’s Cold War decision about when first to nonstate actors, hybrid warfare, nuclear and conventional warfare, hybrid warfare, respond with nuclear weapons to a Russian conventional attack, ambiguity sur- space warfare, and cyberwar (Lindsay and Gartzke, 2016; Manzo, 2011; Scouras, rounding intervention and escalation thresholds can actually be useful and can Smyth, and Mahnken, 2014). The DoD definition is used in this Perspective. contribute to crisis stability. The point here is that the deterrer must be able to 36 The concepts of theescalation threshold, escalation ladder, and escalation path are prevent aggressor attempts to manipulate, blur, and/or compromise thresholds at defined by Kahn (1965, p. 37f). which the deterrer has decided that there is value in having a clear “red line.”

29 37 It is worth noting that this escalation scenario is—by design—relatively mild 49 Temporary “mesh networks” created by high-altitude drones launched from and mostly lies at the low end of the multistep escalation ladder elaborated by across domains from submarines, surface combatants, or aircraft are another, Herman Kahn. lower-cost (albeit temporary) expedient by which to deny the opponent the advantages of attacking satellite communications capacity in times of crisis or war. 38 Davis points out that the discussion that follows resembles work in the 1980s in I am grateful to RAND colleague Zev Winkelman for this suggestion. See also support of what Fred Iklé and others referred to as “horizontal escalation.” See, for Delpech (2012, p. 150). While a seabed fiber optic network would be vulnerable example, Miller and Van Evera (2014, p. 63f) who, in turn, cite Lehman (1981) to opponent attack, its “connectedness” would cause it to decay gracefully under and Iklé (1982) (Davis communication with the author). directed attack. Back-of-the-envelope calculations suggest that such a network 39 For the purposes of this exploratory, concept paper, the applicability to the given might be put in place at a cost of less than $20 billion. This does not appear to be domain or subarea of each of the classical assumptions and underlying enablers an expensive diversification or insurance policy when compared with the possibil- of successful deterrence was reviewed and subjectively scored “low,” “medium,” ity of the catastrophic failure of space-based military communications. or “high.” Given the time and the space, a more nuanced methodology could 50 “Information Space—The sphere of activity connected to forming, creating, certainly have been applied. While this initial analysis is hardly sophisticated, its transforming, transmitting, using and storing information and that has an influ- results are more than just impressionistic. ence amongst other things on individual and social consciousness, the informa- 40 Prior to the latter Chinese demonstration, tacit agreement was thought to exist tion infrastructure and information itself” (Ministry of Defense of the Russian that the GEO in which key satellites are placed constituted a “sanctuary” that Federation, 2011, p. 5). See also Richard Clarke’s discussion of the Chinese would not be attacked (Hitchens and Johnson-Freese, 2016, p. 43; G. H. Snyder, People’s Liberation Army’s concepts of networkization (wangluohua) and informa- 1961, p. 134). tion dominance (zhixinxiquan) in Clarke and Knake (2010).

41 China may, on the other hand, have reached a sober assessment that such an 51 An example of a protected information space is the organizations, comput- attack is one of the few means by which it can counter U.S. conventional military ers, and equipment comprising the “Great Firewall.” The Great Firewall project superiority at the outset of a conflict (Clarke and Knake, 2010). operated by the Bureau of Public Information and Network Security Supervision of the Chinese Ministry of Public Security is a complex of legislation, regula- 42 The Union of Concerned Scientists lists a total of 576 U.S. satellites in orbit tions, and technologies permitting authorities to surveil and censor the Internet versus 181 for China and 140 for Russia. For the condition of Russia’s network of in China. It allows authorities to establish control over the Chinese domestic early warning satellites, see Quinlivan and Oliker (2011, p. 42). information space by a variety of means that include blocking access to websites, 43 See Hitchens and Johnson-Freese (2016). social media, mobile applications, and virtual private networks and by throttling international Internet backbone access rates (Dou, 2015). 44 Delpech, 2012, p. 147. 52 For more than a decade, both public and private entities have been beaming 45 Thenuclear kill chain consists of the various actions and equipment required to television news programming by satellite into Iran. The Iranian authorities have execute a successful attack using nuclear weapons. made repeated attempts to jam such signals. However, these Iranian government 46 Forrest Morgan identifies major escalation thresholds (1) at the border between efforts have mostly been futile because satellite television signals are more difficult reversible and destructive attacks, (2) at the border between destructive attacks to jam than terrestrial radio broadcasts (source: communication with Middle that cause debris fields in orbit and those that do not, and (3) at the point where a East Broadcast Networks management). Satellites launched to provide broad-area nuclear weapon is detonated in space (F. E. Morgan, 2010). television coverage into opponent information spaces might, of course, be prime targets for adversary attack. One method by which to thwart such attacks, which 47 For an explanation of the potential for catastrophic failure when networks are might be termed snuggling, could be to position U.S. satellites so close in GEO to subjected to “directed attack,” see Newman (2010, p. 592f). opponent satellites fulfilling the same function that debris from an attack on the 48 As defined previously, crisis stability is “a measure of countries’ incentives not U.S. satellite would cause damage to or destroy the adversary’s satellite. to preempt in a crisis, that is, not to attack first in order to beat the attack of the 53 NATO Allied Command Transformation, 2011. enemy” (F. E. Morgan, 2010, pp. 1–2, citing Charles Glaser).

30 54 Ganor elaborates an escalating spectrum of deterrent measures that might be critical portion of a community is immunized against a contagious disease, most employed either unilaterally or collectively against state sponsors of proxy terrorist members of the community are protected against that disease because there is organizations (Ganor, 2005, p. 79f). little opportunity for an outbreak. . . . This is known as ‘[community or herd] immunity’” (Vaccines.gov, 2017). In the context of information warfare, inocula- 55 Ablodia, 1993; Asmus, 2010; Blackwill and Harris, 2016; Cohen and Hamilton, tion involves protecting the general public against information warfare tactics by 2011; Cornell and Starr, 2009; Kahn, 1965. providing them with knowledge of opponent tactics and the ability to detect and 56 Kahn, 1965. reject false narratives and disinformation deliberately spread on social media and other broadcast platforms by opposing forces. 57 The state-owned economic entities in question frequently enjoy unrestricted and unreciprocated access to Western financial, raw material, consumer, and industrial 62 Historically, the U.S. Army’s 173rd Airborne Brigade, based in Vicenza, Italy, markets. For greater detail, see Blackwill and Harris (2016). and the U.S. Army’s 82nd and 101st Airborne Divisions stationed at Fort Bragg, North Carolina, and Fort Campbell, Kentucky, have served as QRFs of this kind 58 The key enablers that are present are the ability to deter directly, the ability to for the United States. effect extended deterrence, the low probability of strategic surprise, the presence or availability of escalation thresholds, and a high attacker-to-target ratio. 63 Libicki, 2009. 59 On April 14, 2014, President Barack Obama reaffirmed that U.S. treaty obliga- 64 Treverton and Jones, 2005, pp. 9–10. tions to Japan extended to the Senkaku (Pinnacle) Islands controlled by Japan 65 “Terrorism is a form of violent struggle in which violence is deliberately used and located in the East China Sea. Seven months later, in a climb-down, China against civilian targets in order to achieve political goals” (Ganor, 2005, p. 17f). agreed to open negotiations with Japan over the Islands (Eilperin, 2014; Perlez, 2014). On January 22, 2013, in connection with its dispute with China over the 66 According to the Institute for Economics and Peace, in 2017, fatalities caused Scarborough Shoal, Manila served Beijing with a Notification and Statement of by terrorist attacks on developed Organisation of Economic Co-operation and Claim under Annex II to the UN Convention on the Law of the Sea concerning Development nations (most of them U.S. allies) hit a 16-year high (Bond and its “nine-dashed line” claims at the International Tribunal on the Law of the Sea. Wisniewska, 2017). On October 30, 2015, in a blow to China, the Permanent Court of Arbitration 67 “‘Hardening,’ of which the typical air raid shelter is an example . . . involves put- in the Hague ruled that it had jurisdiction over Manila’s suit. Fearing a Chinese ting a shield between the targets to be protected, whether human or inanimate, act of preemption, President Obama warned Chinese President Xi Jinping at the and the bursting bomb” (Brodie, 1959, p. 210). March 2016 Washington Nuclear Security Summit that if Beijing built on Scar- borough Shoal, there would be a U.S. reaction. To reinforce the point, four U.S. 68 For a useful framework for formulating deterrence policy against terrorist orga- A-10 “Warthog” ground attack aircraft overflew Scarborough Shoal on April 21, nizations, see Ganor (2005, Figure 4.5) and the discussion that follows. 2016. After the Hague issued a sweeping ruling against China’s claims and after 69 conciliatory actions by Philippine President Rodrigo Duterte, China subsequently See Trager and Zagorcheva (2005), Kroenig and Pavel (2012, p. 30), and withdrew its vessels from the area (Cooper and Douglas, 2016; Perlez, 2012, 2015, Abrams (2014, p. 3). See also P. K. Davis and Jenkins (2002). 2016). 70 This is an allusion to the concepts of clan solidarity ˲Δ( ͉ϴ˶Βμ˴ ϋ˴ ) and the blood feud ˸ ˴ 60 “Strategic Deception and Disinformation (ɋɬɪɚɬɟɝɢɱɟɫɤɚɹɦɚɫɤɢɪɨɜɤɚɢ (έ˲ ΄Λ) or lex talionis. See Bruce Malina’s discussion of “Mediterranean persons” ɞɟɡɢɧɮɨɪɦɚɰɢɹ)—a form of strategic support organized and executed for the in Rohrbaugh (1996, Chapter 2) and his discussion of collective identity in the purpose of confusing the opponent about the composition, condition and location “Mediterranean cultural continent” in Malina (2001). The United States has, in of the armed forces, their potential level of supply and combat readiness, about the fact, employed collective punishment tactics in the past. The United States fire- location of strategic bases and their protection, about military construction plans, bombed German and Japanese cities during World War II, killing up to 135,000 the intentions and decisions of the military-political leadership and about strategic in Dresden alone. The cities of Hiroshima and Nagasaki were razed by two U.S. plans” [emphasis added] (Rogozin, 2011, Article 5.92). atomic bombs, killing up to 146,000. The United States also employed napalm against villages during the Vietnam War. 61 The classic article on inoculation effects—research spurred by the brainwashing 71 of U.S. prisoners of war during the Korean War—is McGuire (1961). “When a Ganor, 2005, p. 205.

31 72 At the lowest end of the range of options, such interdiction might involve reduc- Turkey and Ukraine on which Georgia is heavily dependent. Because countries, ing U.S. payments to such organizations as the Palestinian National Authority such as North Korea, have groups of hackers based in other countries (China [in (PNA) by the value of the pensions and other compensation paid to families and Dandong and Sunyang] and Malaysia), cutting off countries’ Internet access in the cost of campaigns promoting acts of terror. In its most extreme form, it might times of crises is not a cure-all (Clarke and Knake, 2010, p. 27; Cohen and Ham- involve ceasing payments to the PNA entirely. The extreme option is, in all likeli- ilton, 2011, p. 45). hood, an unrealistic one. The PNA does play a significant representative role and 82 J. S. Davis et al., 2016. provides services and governance that Israeli authorities might encounter great difficulty delivering in the PNA’s absence. 83 The market would price companies that fail the test differently than those that do not, thereby providing an incentive for greater cybersecurity. 73 See the discussion of collective punishment in Ganor (2005, p. 203ff). 84 Legislation could establish clearer corporate legal liability for the compromise 74 Ganor, 2005, p. 212. of PII, personal health information, and IPR stored on computer systems or 75 Cyberspace is defined as “all of the computer networks in the world and every- “Internet of Things” devices. Just as in the case of household appliances, in order thing they connect and control. It’s not just the Internet” (Clarke and Knake, for companies maintaining such systems or producing Internet of Things equip- 2010, p. 70). The following section has benefited from formal and informal com- ment to receive liability insurance coverage, national testing laboratories could munications with Ted Schlein, Kevin Mandia, Adam Ghetti, Oren Falkowitz, certify their systems as being compliant with a private-sector–designed security Frank Cilluffo, Scott Charney, Carol Haave, Christopher Ford, David Benson, standard, such as the R3D standard (robust by design, robust by default, and and James Farwell. Documents consulted include Delpech (2012, Chapter 7), robust by deployment). Companies operating equipment receiving such certifica- DoD (2011, 2014, and 2015), Farwell and Arkelian (2014), Nye and Joseph tion would, presumably, enjoy a pricing premium, thereby creating a market signal (2017), Slayton (2017), National Institute of Standards and Technology (2014), that encourages improved cybersecurity. The market would likely reflect certified and Defense Science Board (2017). products’ pricing premium in companies’ stock prices, providing a further positive market incentive to adopt better cybersecurity practices. 76 The following words of caution written nearly 60 years earlier about the advent of nuclear weapons by Bernard Brodie may be worth bearing in mind: “The 85 A monopsony is the opposite of a monopoly. In a monopoly, the seller is the sole [military] bias towards the offensive creates special problems in any technologi- supplier of a product. In a monopsony, the consumer is the sole consumer of a cally new situation where there is little or no relevant war experience to help one product. A consumer with monopsony power can dictate prices and product char- to reach a balanced judgment” (Brodie, 1959, p. 175). See also Clarke’s discussion acteristics to sellers and thereby shape the structure of a market; one consumer of the U.S. National Military Strategy for Cyber Operations in Clarke and Knake that is much larger than the others in the market can have inordinate influence (2010, pp. 44f, 115). over sellers.

77 This threat includes, in its most extreme form, attacks on critical national 86 The Social Security Administration (2000) gives examples of how biometric infrastructure. information has been used to prevent dozens of millions of dollars in fraud. Train- ing of computer users consistently ranks highly amongst the methods by which to 78 This threat includes the threat of cyberattack on computer systems controlling protect against attempts to compromise computer system security by use of social U.S. critical infrastructure, vividly described in a scenario in Clarke and Knake engineering. See, for example, Olavsrud (2010). (2010, p. 63f). 87 Libicki (2009); see also J. S. Davis et al. (2016). See Gorman, Cole, and Dreazen 79 See, for instance, Ravich (2017). (2009) for the importance to national security of protecting private-sector IPR 80 Brodie, 1959, p. 175. See also Clarke’s discussion of the U.S. National Military from cyber theft. When digesting Libicki’s assertion that the offensive dominates Strategy for Cyber Operations in Clarke and Knake (2010, p. 45). in cyberspace, the following words of caution written nearly 60 years earlier about the advent of nuclear weapons by Brodie may—once again—be worth bearing in 81 During the 2008 Russo-Georgian war, Internet connections between Russia mind: “The [military] bias towards the offensive creates special problems in any and Georgia were shut down to thwart Russian cyberattacks. Russia responded by technologically new situation where there is little or no relevant war experience to conducting cyberattacks via connections through third countries (China, Canada, help one to reach a balanced judgment” (Brodie, 1959, p. 175). Turkey, and Estonia) and with cyberattacks on Internet backbone connections to

32 88 Farwell and Arkelian, 2014. 89 I am grateful to David Benson, currently of the Air University, for bringing the legal instrument of the international convention as a vehicle for collective cyber defense to my attention. 90 Davis describes a simulation conducted at RAND in the late 1980s that drew precisely the opposite conclusion. The simulation showed the importance of context, “Red agent” mindsets, and—above all—the time factor in nuclear escalation calculations. Under NATO doctrine at that time, after the outbreak of conventional hostilities with the Warsaw Pact, the allies would continue to fight a conventional war until they were about to lose and would then make limited use of nuclear weapons in order to force the Warsaw Pact to terminate hostilities immediately. Davis concluded that such an effort to reestablish deterrence would likely fail because the Warsaw Pact might feel that it was so close to victory that it might either absorb the pain of “riding out” a limited NATO nuclear attack or escalate with a massive nuclear counterstrike before terminating hostilities. Davis further concluded that much earlier use of nuclear weapons than envisioned by then-current doctrine—i.e., immediate escalation—might be more effective in ensuring immediate deterrence of further hostilities (P. K. Davis, 1989).

91 However, the caveats regarding feasibility, affordability, and maintaining the integrity of U.S. values may, in the end, cause punishment to be the only viable option. 92 For the concept of the partitioning of strategy sets, see Rasmusen (2007).

33 Abbreviations

ASAT antisatellite C3ISR command, control, communication, intelligence, surveillance, and reconnaissance DoD U.S. Department of Defense FFRDC federally funded research and development center GEO geosynchronous Earth orbit HEO highly elliptical orbit IO information operations IPR intellectual property rights ISPK Institute for Security Policy at Kiel University LEO low Earth orbit NATO North Atlantic Treaty Organization NC2 nuclear command and control NDRI National Defense Research Institute PII personally identifiable information PNA Palestinian National Authority QRF quick reaction force R3D robust by design, robust by default, and robust by deployment RSTA reconnaissance, surveillance, targeting, and attack WTO World Trade Organization

34 About This Perspective Additional thanks are due to the associate director of the RAND Center for Asia-Pacific Policy, Scott Harold. Harold arranged for a series of This Perspective examines ways and means by which the United States and structured interviews on cross-domain deterrence in Tokyo with Japanese its allies might meet new challenges in cross-domain deterrence. Cross- experts and decisionmakers. The opportunity to understand the perspec- domain deterrence in four discrete domains or subareas of warfare is exam- tives of this important Asian ally was very helpful. ined: space, hybrid warfare, terrorism, and cyberwarfare. This Perspective Particular thanks go to Jerry Sollinger for his assistance in consolidat- may be of interest to general audiences and to specialists in deterrence ing and editing the text. theory, defense in space, hybrid warfare, information operations, and The author alone bears responsibility for any errors and/or omissions. cyberwarfare. This research was sponsored by the Office of the Secretary of Defense and conducted within the International Security and Defense Policy Center About the Author of the RAND National Defense Research Institute (NDRI), a federally funded King Mallory is a senior researcher at the RAND Corporation’s Boston research and development center (FFRDC) sponsored by the Office of the office. He served as CEO of the Aspen Institute Germany (2007–2013) Secretary of Defense, the Joint Staff, the Unified Combatant Commands, and as the senior adviser to Assistant Secretaries of State for Near Eastern the Navy, the Marine Corps, the defense agencies, and the defense Affairs Bill Burns and David Welch and their (Principal) Deputy Liz Cheney Intelligence Community. from 2002 to 2007. For more information on the RAND International Security and Defense Policy Center, see http://www.rand.org/nsrd/ndri/centers/isdp.html or contact the director (contact information is provided on the web page). The author would like to thank RAND colleagues Ezra Hecker, Patrick

Orr, Luke Matthews, Joshua Mendelsohn, Zev Winkelman, Bill Marcellino, Limited Print and Electronic Distribution Rights Elizabeth Bodine-Baron, Cynthia Dion-Schwarz, Andrew Scobell, Andrew This document and trademark(s) contained herein are protected by law. This representa- Radin, Paul Davis, Seth Jones, and Jack Riley, as well as former RAND tion of RAND intellectual property is provided for noncommercial use only. Unauthor- president and chief executive officer Jim Thomson, for their assistance in the ized posting of this publication online is prohibited. Permission is given to duplicate this document for personal use only, as long as it is unaltered and complete. Permission is preparation of this paper. required from RAND to reproduce, or reuse in another form, any of our research docu- Special thanks are owed to Professor Joachim Krause of the Institute for ments for commercial use. For information on reprint and linking permissions, please visit Security Policy at Kiel University (ISPK). RAND partnered with ISPK to hold www.rand.org/pubs/permissions.html. a one-day workshop in Berlin on cross-domain deterrence hosted by ISPK. The RAND Corporation is a research organization that develops solutions to public Forty military officers and national security experts from the United States policy challenges to help make communities throughout the world safer and more secure, healthier and more prosperous. RAND is nonprofit, nonpartisan, and committed to the and several European countries (including Finland, Sweden, Germany, the public interest. United Kingdom, and France) attended. The workshop provided valuable RAND’s publications do not necessarily reflect the opinions of its research clients and insights into how a doctrine of cross-domain deterrence might be imple- sponsors. R® is a registered trademark. mented in Europe. For more information on this publication, visit www.rand.org/t/PE259.

PE-259-OSD (2018)